{ "title": "Efficiency in LLM Multi-Agent Systems: A Survey", "generatedFrom": "Active manuscript citations and curated project-page additions", "sourceSHA256": "2c168205341201922f8013a925d74d3e4f152bc2594ba8e73a4d187cb61b3c98", "classification": "Topic tags indicate relevance. Manuscript-derived tags retain citation placement; project-page additions are curated separately. Evidence-table methods retain the manuscript's primary family.", "stats": { "cited": 256, "bibEntries": 272, "evidenceMethods": 29, "bySection": { "Background": 17, "Collaboration Boundary": 10, "Runtime": 116, "Topology": 27, "Optimization": 31, "Synthesis": 13, "Evaluation": 68, "Introduction": 31, "Open Problems": 47 }, "collected": 257, "supplemental": 1 }, "papers": [ { "key": "1980_smith_contract_net", "title": "The Contract Net Protocol: High-Level Communication and Control in a Distributed Problem Solver", "authors": "Reid G. Smith", "year": 1980, "url": "https://doi.org/10.1109/TC.1980.1675516", "venue": "IEEE Transactions on Computers", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "1994_littman_markov_games", "title": "Markov Games as a Framework for Multi-Agent Reinforcement Learning", "authors": "Michael L. Littman", "year": 1994, "url": "https://doi.org/10.1016/B978-1-55860-335-6.50027-1", "venue": "11th International Conference on Machine Learning (ICML-94)", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "1995_rao_bdi", "title": "BDI Agents: From Theory to Practice", "authors": "Anand S. Rao and Michael P. Georgeff", "year": 1995, "url": "https://cdn.aaai.org/ICMAS/1995/ICMAS95-042.pdf", "venue": "First International Conference on Multi-Agent Systems (ICMAS-95)", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2002_bernstein_decpomdp", "title": "The Complexity of Decentralized Control of Markov Decision Processes", "authors": "Daniel S. Bernstein and Robert Givan and Neil Immerman and Shlomo Zilberstein", "year": 2002, "url": "https://doi.org/10.1287/moor.27.4.819.297", "venue": "Mathematics of Operations Research", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2017_lowe_maddpg", "title": "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments", "authors": "Ryan Lowe and Yi Wu and Aviv Tamar and Jean Harb and Pieter Abbeel and Igor Mordatch", "year": 2017, "url": "https://proceedings.neurips.cc/paper_files/paper/2017/hash/68a9750337a418a86fe06c1991a1d64c-Abstract.html", "venue": "Conference on Neural Information Processing Systems (NeurIPS-17)", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2018_rashid_qmix", "title": "QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning", "authors": "Tabish Rashid and Mikayel Samvelyan and Christian Schroeder de Witt and Gregory Farquhar and Jakob Foerster and Shimon Whiteson", "year": 2018, "url": "https://proceedings.mlr.press/v80/rashid18a.html", "venue": "35th International Conference on Machine Learning (ICML-18)", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2022_khot_decomposed_prompting", "title": "Decomposed Prompting: A Modular Approach for Solving Complex Tasks", "authors": "Tushar Khot and Harsh Trivedi and Matthew Finlayson and Yao Fu and Kyle Richardson and Peter Clark and Ashish Sabharwal", "year": 2022, "url": "https://arxiv.org/abs/2210.02406", "venue": "International Conference on Learning Representations (ICLR-23)", "categories": [ "Collaboration Boundary" ], "subfamilies": [], "sourceFiles": [ "sections/3_boundaries.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_holt_l2mac", "title": "L2MAC: Large Language Model Automatic Computer for Extensive Code Generation", "authors": "Samuel Holt and Max Ruiz Luyten and Mihaela van der Schaar", "year": 2023, "url": "https://arxiv.org/abs/2310.02003", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_hu_commformer", "title": "Learning Multi-Agent Communication from Graph Modeling Perspective", "authors": "Shengchao Hu and Li Shen and Ya Zhang and Dacheng Tao", "year": 2024, "url": "https://arxiv.org/abs/2405.08550", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Topology" ], "subfamilies": [], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_hua_waragent", "title": "War and Peace (WarAgent): Large Language Model-based Multi-Agent Simulation of World Wars", "authors": "Wenyue Hua and Lizhou Fan and Lingyao Li and Kai Mei and Jianchao Ji and Yingqiang Ge and Libby Hemphill and Yongfeng Zhang", "year": 2023, "url": "https://arxiv.org/abs/2311.17227", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_jiang_llm_blender", "title": "LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion", "authors": "Dongfu Jiang and Xiang Ren and Bill Yuchen Lin", "year": 2023, "url": "https://arxiv.org/abs/2306.02561", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-23)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_liang_mad", "title": "Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate", "authors": "Tian Liang and Zhiwei He and Wenxiang Jiao and Xing Wang and Yan Wang and Rui Wang and Yujiu Yang and Shuming Shi and Zhaopeng Tu", "year": 2023, "url": "https://arxiv.org/abs/2305.19118", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_packer_memgpt", "title": "MemGPT: Towards LLMs as Operating Systems", "authors": "Charles Packer and Sarah Wooders and Kevin Lin and Vivian Fang and Shishir G. Patil and Ion Stoica and Joseph E. Gonzalez", "year": 2023, "url": "https://arxiv.org/abs/2310.08560", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_pryzant_protegi", "title": "Automatic prompt optimization with “gradient descent” and beam search", "authors": "Reid Pryzant and Dan Iter and Jerry Li and Yin Lee and Chenguang Zhu and Michael Zeng", "year": 2023, "url": "https://aclanthology.org/2023.emnlp-main.494/", "venue": "Proceedings of the 2023 conference on empirical methods in natural language processing", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "https://aclanthology.org/2023.emnlp-main.494/" }, { "key": "2023_talebirad_multiagent_collaboration_harnessing", "title": "Multi-Agent Collaboration: Harnessing the Power of Intelligent LLM Agents", "authors": "Yashar Talebirad and Amirhossein Nadiri", "year": 2023, "url": "https://arxiv.org/abs/2306.03314", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_yin_exchange_of_thought", "title": "Exchange-of-Thought: Enhancing Large Language Model Capabilities through Cross-Model Communication", "authors": "Zhangyue Yin and Qiushi Sun and Cheng Chang and Qipeng Guo and Junqi Dai and Xuanjing Huang and Xipeng Qiu", "year": 2023, "url": "https://arxiv.org/abs/2312.01823", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-23)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_zhao_expel", "title": "ExpeL: LLM Agents Are Experiential Learners", "authors": "Andrew Zhao and Daniel Huang and Quentin Xu and Matthieu Lin and Yong-Jin Liu and Gao Huang", "year": 2023, "url": "https://arxiv.org/abs/2308.10144", "venue": "38th Annual AAAI Conference on Artificial Intelligence (AAAI-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_zhong_memorybank", "title": "MemoryBank: Enhancing Large Language Models with Long-Term Memory", "authors": "Wanjun Zhong and Lianghong Guo and Qiqi Gao and He Ye and Yanlin Wang", "year": 2023, "url": "https://arxiv.org/abs/2305.10250", "venue": "38th Annual AAAI Conference on Artificial Intelligence (AAAI-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2023_zhou_agents", "title": "Agents: An Open-source Framework for Autonomous Language Agents", "authors": "Wangchunshu Zhou and Yuchen Eleanor Jiang and Long Li and Jialong Wu and Tiannan Wang and Shi Qiu and Jintian Zhang and Jing Chen and Ruipu Wu and Shuai Wang and Shiding Zhu and Jiyu Chen and Wentao Zhang and Xiangru Tang and Ningyu Zhang and Huajun Chen and Peng Cui and Mrinmaya Sachan", "year": 2023, "url": "https://arxiv.org/abs/2309.07870", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2024_chen_optima", "title": "Optima: Optimizing Effectiveness and Efficiency for LLM-Based Multi-Agent System", "authors": "Weize Chen and Jiarui Yuan and Chen Qian and Cheng Yang and Zhiyuan Liu and Maosong Sun", "year": 2024, "url": "https://arxiv.org/abs/2410.08115", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-25)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Policy Learning" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex" ], "method": "OPTIMA", "evidence": { "method": "OPTIMA", "family": "Optimization", "subfamily": "Policy", "update": "training", "effect": "message tokens", "boundary": "quality–length trade-off", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2024_chen_smurfs", "title": "Smurfs: Multi-Agent System using Context-Efficient DFSDT for Tool Planning", "authors": "Junzhi Chen and Juhao Liang and Benyou Wang", "year": 2024, "url": "https://arxiv.org/abs/2405.05955", "venue": "Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL-25)", "categories": [ "Runtime", "Synthesis" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/6_synthesis.tex" ], "method": "Smurfs", "evidence": { "method": "Smurfs", "family": "Runtime", "subfamily": "Routing", "update": "runtime", "effect": "tokens and pass rate", "boundary": "tool retries remain", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2024_collaborative_rag", "title": "A Collaborative Multi-Agent Approach to Retrieval-Augmented Generation Across Diverse Data", "authors": "Aniruddha Salve and Saba Attar and Mahesh Deshmukh and Sayali Shivpuje and Arnab Mitra Utsab", "year": 2024, "url": "https://arxiv.org/abs/2412.05838", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication", "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2024_gandhi_budgetmlagent", "title": "BudgetMLAgent: A Cost-Effective LLM Multi-Agent system for Automating Machine Learning Tasks", "authors": "Shubham Gandhi and Manasi Patwardhan and Lovekesh Vig and Gautam Shroff", "year": 2024, "url": "https://arxiv.org/abs/2411.07464", "venue": "International Conference on AI-ML Systems (AIMLSystems-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2024_guo_stabletoolbench", "title": "StableToolBench: Towards Stable Large-Scale Benchmarking on Tool Learning of Large Language Models", "authors": "Zhicheng Guo and Sijie Cheng and Hao Wang and Shihao Liang and Yujia Qin and Peng Li and Zhiyuan Liu and Maosong Sun and Yang Liu", "year": 2024, "url": "https://aclanthology.org/2024.findings-acl.664/", "venue": "Findings of the Association for Computational Linguistics: ACL 2024 (ACL Findings-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2024_islam_mapcoder", "title": "MapCoder: Multi-Agent Code Generation for Competitive Problem Solving", "authors": "Md. Ashraful Islam and Mohammed Eunus Ali and Md Rizwan Parvez", "year": 2024, "url": "https://arxiv.org/abs/2405.11403", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Runtime", "Synthesis" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/6_synthesis.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2024_zhang_machinesom", "title": "Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View", "authors": "Jintian Zhang and Xin Xu and Ningyu Zhang and Ruibo Liu and Bryan Hooi and Shumin Deng", "year": 2023, "url": "https://arxiv.org/abs/2310.02124", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2024_zhou_llm_mapreduce", "title": "LLM×MapReduce: Simplified Long-Sequence Processing using Large Language Models", "authors": "Zihan Zhou and Chong Li and Xinyi Chen and Shuo Wang and Yu Chao and Zhili Li and Haoyu Wang and Rongqiao An and Qi Shi and Zhixing Tan and Xu Han and Xiaodong Shi and Zhiyuan Liu and Maosong Sun", "year": 2024, "url": "https://arxiv.org/abs/2410.09342", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_besrour_ragenta", "title": "RAGentA: Multi-Agent Retrieval-Augmented Generation for Attributed Question Answering", "authors": "Ines Besrour and Jingbo He and Tobias Schreieder and Michael Färber", "year": 2025, "url": "https://arxiv.org/abs/2506.16988", "venue": "ACM SIGIR Conference on Research and Development in Information Retrieval, LiveRAG Challenge (SIGIR LiveRAG-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_cemri_mast", "title": "Why Do Multi-Agent LLM Systems Fail?", "authors": "Mert Cemri and Melissa Z. Pan and Shuyi Yang and Lakshya A. Agrawal and Bhavya Chopra and Rishabh Tiwari and Kurt Keutzer and Aditya Parameswaran and Dan Klein and Kannan Ramchandran and Matei Zaharia and Joseph E. Gonzalez and Ion Stoica", "year": 2025, "url": "https://arxiv.org/abs/2503.13657", "venue": "Conference on Neural Information Processing Systems (NeurIPS-25)", "categories": [ "Runtime", "Evaluation" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_chang_anp_whitepaper", "title": "Agent Network Protocol Technical White Paper", "authors": "Gaowei Chang and Eidan Lin and Chengxuan Yuan and Rizhao Cai and Binbin Chen and Xuan Xie and Yin Zhang", "year": 2025, "url": "https://arxiv.org/abs/2508.00007", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_chang_main_rag", "title": "MAIN-RAG: Multi-Agent Filtering Retrieval-Augmented Generation", "authors": "Chia-Yuan Chang and Zhimeng Jiang and Vineeth Rakesh and Menghai Pan and Chin-Chia Michael Yeh and Guanchu Wang and Mingzhi Hu and Zhichao Xu and Yan Zheng and Mahashweta Das and Na Zou", "year": 2024, "url": "https://arxiv.org/abs/2501.00332", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-25)", "categories": [ "Runtime", "Collaboration Boundary" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/3_boundaries.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_deng_agentic_services_computing", "title": "Agentic Services Computing", "authors": "Shuiguang Deng and Hailiang Zhao and Ziqi Wang and Wenzhuo Qian and Xiang Ao and Guanjie Cheng and Jianwei Yin and Albert Y. Zomaya and Schahram Dustdar", "year": 2025, "url": "https://arxiv.org/abs/2509.24380", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_ehtesham_agent_interop_survey", "title": "A survey of agent interoperability protocols: Model Context Protocol (MCP), Agent Communication Protocol (ACP), Agent-to-Agent Protocol (A2A), and Agent Network Protocol (ANP)", "authors": "Abul Ehtesham and Aditi Singh and Gaurav Kumar Gupta and Saket Kumar", "year": 2025, "url": "https://arxiv.org/abs/2505.02279", "venue": "arXiv preprint", "categories": [ "Topology", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_fu_c2c", "title": "Cache-to-Cache: Direct Semantic Communication Between Large Language Models", "authors": "Tianyu Fu and Zihan Min and Hanling Zhang and Jichao Yan and Guohao Dai and Wanli Ouyang and Yu Wang", "year": 2025, "url": "https://arxiv.org/abs/2510.03215", "venue": "International Conference on Learning Representations (ICLR-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_han_mapgd", "title": "MAPGD: Multi-Agent Prompt Gradient Descent for Collaborative Prompt Optimization", "authors": "Yichen Han and Yuhang Han and Siteng Huang and Guanyu Liu and Zhengpeng Zhou and Bojun Liu and Yujia Zhang and Isaac N Shi and Lewei He and Tianyu Shi", "year": 2025, "url": "https://arxiv.org/abs/2509.11361", "venue": "arXiv preprint", "categories": [ "Optimization", "Synthesis" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex", "sections/6_synthesis.tex" ], "method": "MAPGD", "evidence": { "method": "MAPGD", "family": "Optimization", "subfamily": "Prompt", "update": "design", "effect": "search tokens, calls, and F1", "boundary": "wall time increases", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Offline cost included (O)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_jiang_gtd", "title": "Dynamic Generation of Multi-LLM Agents Communication Topologies with Graph Diffusion Models", "authors": "Eric Hanchen Jiang and Mengting Li and Guancheng Wan and Sophia Yin and Yuchen Wu and Xiao Liang and Xinfeng Li and Yizhou Sun and Wei Wang and Kai-Wei Chang and Ying Nian Wu", "year": 2025, "url": "https://arxiv.org/abs/2510.07799", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-26)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "GTD", "evidence": { "method": "GTD", "family": "Topology", "subfamily": "Construction", "update": "offline + query", "effect": "communication", "boundary": "diffusion and proxy cost", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_jin_corl", "title": "Controlling Performance and Budget of a Centralized Multi-agent LLM System with Reinforcement Learning", "authors": "Bowen Jin and TJ Collins and Donghan Yu and Mert Cemri and Shenao Zhang and Mengyu Li and Jay Tang and Tian Qin and Zhiyang Xu and Jiarui Lu and Guoli Yin and Jiawei Han and Zirui Wang", "year": 2025, "url": "https://arxiv.org/abs/2511.02755", "venue": "arXiv preprint", "categories": [ "Runtime", "Optimization" ], "subfamilies": [ "Optimization / Policy Learning", "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/4_3_optimization.tex" ], "method": "CORL", "evidence": { "method": "CORL", "family": "Optimization", "subfamily": "Policy", "update": "training", "effect": "calls and dollars", "boundary": "rollout cost", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_kang_self_certainty_bon", "title": "Scalable Best-of-N Selection for Large Language Models via Self-Certainty", "authors": "Zhewei Kang and Xuandong Zhao and Dawn Song", "year": 2025, "url": "https://arxiv.org/abs/2502.18581", "venue": "Conference on Neural Information Processing Systems (NeurIPS-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_kriuk_qkvcomm", "title": "Q-KVComm: Efficient Multi-Agent Communication Via Adaptive KV Cache Compression", "authors": "Boris Kriuk and Logic Ng", "year": 2025, "url": "https://arxiv.org/abs/2512.17914", "venue": "International Conference on Multi-Agent Systems for Collaborative Intelligence (ICMSCI-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_leong_amas", "title": "AMAS: Adaptively Determining Communication Topology for LLM-based Multi-Agent System", "authors": "Hui Yi Leong and Yuheng Li and Yuqing Wu and Wenwen Ouyang and Wei Zhu and Jiechao Gao and Wei Han", "year": 2025, "url": "https://arxiv.org/abs/2510.01617", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-25)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_mou_ecolang", "title": "EcoLANG: Efficient and Effective Agent Communication Language Induction for Social Simulation", "authors": "Xinyi Mou and Chen Qian and Wei Liu and Ling Yan and Yao Hu and Xuanjing Huang and Zhongyu Wei", "year": 2025, "url": "https://aclanthology.org/2025.findings-emnlp.284/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2025 (EMNLP Findings-25)", "categories": [ "Runtime", "Synthesis" ], "subfamilies": [ "Runtime / Communication", "Runtime / Routing" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/6_synthesis.tex" ], "method": "EcoLANG", "evidence": { "method": "EcoLANG", "family": "Runtime", "subfamily": "Communication", "update": "design + runtime", "effect": "response and total tokens", "boundary": "induction cost excluded", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_nguyen_ma_rag", "title": "MA-RAG: Multi-Agent Retrieval-Augmented Generation via Collaborative Chain-of-Thought Reasoning", "authors": "Thang Nguyen and Peter Chin and Yu-Wing Tai", "year": 2025, "url": "https://arxiv.org/abs/2505.20096", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_niu_flow", "title": "Flow: Modularized Agentic Workflow Automation", "authors": "Boye Niu and Yiliao Song and Kai Lian and Yifan Shen and Yu Yao and Kun Zhang and Tongliang Liu", "year": 2025, "url": "https://arxiv.org/abs/2501.07834", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_ren_agentradio", "title": "AgentRadio: Passive Awareness for Long-Horizon Multi-Agent Collaboration", "authors": "Xinxing Ren and Qianbo Zang and Ziyan Wang and Caelum Forder and Suman Deb and Peter Carroll and Zekun Guo", "year": 2026, "url": "https://arxiv.org/abs/2607.28430", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex" ], "method": "AgentRadio", "evidence": { "method": "AgentRadio", "family": "Runtime", "subfamily": "Scheduling", "update": "runtime/step", "effect": "task quality and API cost", "boundary": "wall time not reported", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_rezazadeh_collaborative_memory", "title": "Collaborative Memory: Multi-User Memory Sharing in LLM Agents with Dynamic Access Control", "authors": "Alireza Rezazadeh and Zichao Li and Ange Lou and Yuying Zhao and Wei Wei and Yujia Bao", "year": 2025, "url": "https://arxiv.org/abs/2505.18279", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_shen_masc", "title": "Metacognitive Self-Correction for Multi-Agent System via Prototype-Guided Next-Execution Reconstruction", "authors": "Xu Shen and Qi Zhang and Song Wang and Zhen Tan and Xinyu Zhao and Laura Yao and Vaishnav Tadiparthi and Hossein Nourkhiz Mahjoub and Ehsan Moradi Pari and Kwonjoon Lee and Tianlong Chen", "year": 2025, "url": "https://arxiv.org/abs/2510.14319", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_shen_relational_priors_as", "title": "Relational Priors as Convergence Pressure in LLM-Based Multi-Agent Systems", "authors": "Ming Shen and Chao Shang and Sadat Shahriar and Devang Kulshreshtha and Yi Zhang and Sandesh Swamy and Yanjun Qi", "year": 2026, "url": "https://arxiv.org/abs/2608.03239", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_singh_agentic_rag_survey", "title": "Agentic Retrieval-Augmented Generation: A Survey on Agentic RAG", "authors": "Aditi Singh and Abul Ehtesham and Saket Kumar and Tala Talaei Khoei and Athanasios V. Vasilakos", "year": 2025, "url": "https://arxiv.org/abs/2501.09136", "venue": "arXiv preprint", "categories": [ "Runtime", "Introduction" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/1_intro.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_tang_agent_kb", "title": "Agent KB: Leveraging Cross-Domain Experience for Agentic Problem Solving", "authors": "Xiangru Tang and Tianrui Qin and Tianhao Peng and Ziyang Zhou and Daniel Shao and Tingting Du and Xinming Wei and Peng Xia and Fang Wu and He Zhu and Ge Zhang and Jiaheng Liu and Xingyao Wang and Sirui Hong and Chenglin Wu and Hao Cheng and Chi Wang and Wangchunshu Zhou", "year": 2025, "url": "https://arxiv.org/abs/2507.06229", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_wang_agenttaxo", "title": "AgentTaxo: Dissecting and Benchmarking Token Distribution of LLM Multi-Agent Systems", "authors": "Qian Wang and Zhenheng Tang and Zichen Jiang and Nuo Chen and Tianyu Wang and Bingsheng He", "year": 2025, "url": "https://openreview.net/forum?id=0iLbiYYIpC", "venue": "International Conference on Learning Representations Workshop on Foundation Models in the Wild (ICLR FM-Wild-25)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_wang_masfact", "title": "MasFACT: Continual Multi-Agent Topology Learning via Geometry-Aware Posterior Transfer", "authors": "Xuefei Wang and Jialu Wang and Fengbo Zhang and Yihan Hu and Di Zhang and Yutong Ye and Yikun Ban and Jun Han and Ruijie Wang", "year": 2026, "url": "https://arxiv.org/abs/2605.17361", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Adaptation" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "MasFACT", "evidence": { "method": "MasFACT", "family": "Topology", "subfamily": "Adaptation", "update": "continual + query", "effect": "topology transfer", "boundary": "drift across task streams", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_wang_mirix", "title": "MIRIX: Multi-Agent Memory System for LLM-Based Agents", "authors": "Yu Wang and Xi Chen", "year": 2025, "url": "https://arxiv.org/abs/2507.07957", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_wu_memory_mas_survey", "title": "Memory in LLM-based Multi-agent Systems: Mechanisms, Challenges, and Collective Intelligence", "authors": "Shanglin Wu and Kai Shu", "year": 2025, "url": "https://doi.org/10.36227/techrxiv.176539617.79044553/v1", "venue": "TechRxiv preprint", "categories": [ "Runtime", "Introduction" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/1_intro.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_xu_amem", "title": "A-MEM: Agentic Memory for LLM Agents", "authors": "Wujiang Xu and Zujie Liang and Kai Mei and Hang Gao and Juntao Tan and Yongfeng Zhang", "year": 2025, "url": "https://arxiv.org/abs/2502.12110", "venue": "Conference on Neural Information Processing Systems (NeurIPS-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_xu_divide_conquer_noise", "title": "When Does Divide and Conquer Work for Long Context LLM? A Noise Decomposition Framework", "authors": "Zhen Xu and Shang Zhu and Jue Wang and Junlin Wang and Ben Athiwaratkun and Chi Wang and James Zou and Ce Zhang", "year": 2025, "url": "https://arxiv.org/abs/2506.16411", "venue": "International Conference on Learning Representations (ICLR-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_yang_agentnet", "title": "AgentNet: Decentralized Evolutionary Coordination for LLM-based Multi-Agent Systems", "authors": "Yingxuan Yang and Huacan Chai and Shuai Shao and Yuanyi Song and Siyuan Qi and Renting Rui and Weinan Zhang", "year": 2025, "url": "https://arxiv.org/abs/2504.00587", "venue": "Conference on Neural Information Processing Systems (NeurIPS-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_yang_codeagents", "title": "CodeAgents: A Token-Efficient Framework for Codified Multi-Agent Reasoning in LLMs", "authors": "Bruce Yang and Xinfeng He and Huan Gao and Yifan Cao and Xiaofan Li and David Hsu", "year": 2025, "url": "https://arxiv.org/abs/2507.03254", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_yang_maestro", "title": "Maestro: Learning to Collaborate via Conditional Listwise Policy Optimization for Multi-Agent LLMs", "authors": "Wei Yang and Jiacheng Pang and Shixuan Li and Paul Bogdan and Stephen Tu and Jesse Thomason", "year": 2025, "url": "https://arxiv.org/abs/2511.06134", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_ye_kvcomm", "title": "KVCOMM: Online Cross-context KV-cache Communication for Efficient LLM-based Multi-agent Systems", "authors": "Hancheng Ye and Zhengqi Gao and Mingyuan Ma and Qinsi Wang and Yuzhe Fu and Ming-Yu Chung and Yueqian Lin and Zhijian Liu and Jianyi Zhang and Danyang Zhuo and Yiran Chen", "year": 2025, "url": "https://arxiv.org/abs/2510.12872", "venue": "Conference on Neural Information Processing Systems (NeurIPS-25)", "categories": [ "Runtime", "Synthesis" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/6_synthesis.tex" ], "method": "KVCOMM", "evidence": { "method": "KVCOMM", "family": "Runtime", "subfamily": "State", "update": "runtime", "effect": "KV traffic, prefill", "boundary": "compatibility assumptions", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_yin_empirical_agent_frameworks", "title": "A Comprehensive Empirical Evaluation of Agent Frameworks on Code-centric Software Engineering Tasks", "authors": "Zhuowen Yin and Cuifeng Gao and Chunsong Fan and Wenzhang Yang and Yinxing Xue and Lijun Zhang", "year": 2025, "url": "https://arxiv.org/abs/2511.00872", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_yuen_intrinsic_memory", "title": "Intrinsic Memory Agents: Heterogeneous Multi-Agent LLM Systems through Structured Contextual Memory", "authors": "Sizhe Yuen and Francisco Gomez Medina and Ting Su and Yali Du and Adam J. Sobey", "year": 2025, "url": "https://arxiv.org/abs/2508.08997", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_zeng_s2mad", "title": "S^2-MAD: Breaking the Token Barrier to Enhance Multi-Agent Debate Efficiency", "authors": "Yuting Zeng and Weizhe Huang and Lei Jiang and Tongxuan Liu and Xitai Jin and Chen Tianying Tiana and Jing Li and Xiaohua Xu", "year": 2025, "url": "https://arxiv.org/abs/2502.04790", "venue": "Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex" ], "method": "S2-MAD", "evidence": { "method": "S2-MAD", "family": "Runtime", "subfamily": "Communication", "update": "runtime", "effect": "message tokens", "boundary": "evidence retention", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_zhang_agent-radar", "title": "Enhancing Multi-Agent Communication through Attention Steering with Context Relevance", "authors": "Hongxiang Zhang and Yuan Tian and Tianyi Zhang", "year": 2026, "url": "https://arxiv.org/abs/2605.30136", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_zhang_failureattribution", "title": "Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems", "authors": "Shaokun Zhang and Ming Yin and Jieyu Zhang and Jiale Liu and Zhiguang Han and Jingyang Zhang and Beibin Li and Chi Wang and Huazheng Wang and Yiran Chen and Qingyun Wu", "year": 2025, "url": "https://arxiv.org/abs/2505.00212", "venue": "International Conference on Machine Learning (ICML-25)", "categories": [ "Runtime", "Evaluation" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_zhang_gmemory", "title": "G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems", "authors": "Guibin Zhang and Muxin Fu and Guancheng Wan and Miao Yu and Kun Wang and Shuicheng Yan", "year": 2025, "url": "https://arxiv.org/abs/2506.07398", "venue": "Conference on Neural Information Processing Systems (NeurIPS-25)", "categories": [ "Runtime", "Optimization" ], "subfamilies": [ "Optimization / Continual Learning", "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2025_zhang_metaagent", "title": "MetaAgent: Automatically Constructing Multi-Agent Systems Based on Finite State Machines", "authors": "Yaolun Zhang and Xiaogeng Liu and Chaowei Xiao", "year": 2025, "url": "https://arxiv.org/abs/2507.22606", "venue": "International Conference on Machine Learning (ICML-25)", "categories": [ "Optimization", "Synthesis" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex", "sections/6_synthesis.tex" ], "method": "MetaAgent", "evidence": { "method": "MetaAgent", "family": "Optimization", "subfamily": "Workflow", "update": "task design", "effect": "design and deployment tokens", "boundary": "human-design cost unmeasured", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Offline cost included (O)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2025_zou_latentmas", "title": "Latent Collaboration in Multi-Agent Systems", "authors": "Jiaru Zou and Ruizhong Qiu and Gaotang Li and Xiyuan Yang and Katherine Tieu and Pan Lu and Ke Shen and Hanghang Tong and Yejin Choi and Jingrui He and James Zou and Mengdi Wang and Ling Yang", "year": 2025, "url": "https://arxiv.org/abs/2511.20639", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication", "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_bai_agent_spend_money", "title": "How Do AI Agents Spend Your Money? Analyzing and Predicting Token Consumption in Agentic Coding Tasks", "authors": "Longju Bai and Zhemin Huang and Xingyao Wang and Jiao Sun and Rada Mihalcea and Erik Brynjolfsson and Alex Pentland and Jiaxin Pei", "year": 2026, "url": "https://arxiv.org/abs/2604.22750", "venue": "arXiv preprint", "categories": [ "Runtime", "Evaluation" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_bai_maspromptbench", "title": "MAS-PromptBench: When Does Prompt Optimization Improve Multi-Agent LLM Systems?", "authors": "Juyang Bai and Laixi Shi", "year": 2026, "url": "https://arxiv.org/abs/2606.23664", "venue": "arXiv preprint", "categories": [ "Optimization", "Evaluation" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_3_optimization.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_becker_mint", "title": "Misinformation Propagation in Benign Multi-Agent Systems", "authors": "Jonas Becker and Jan Philip Wahle and Terry Ruas and Bela Gipp", "year": 2026, "url": "https://arxiv.org/abs/2606.16710", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_bertalanic_ringelmann_scaling", "title": "The Ringelmann Effect in Multi-Agent LLM Systems: A Scaling Law for Effective Team Size", "authors": "Blaž Bertalanič and Carolina Fortuna", "year": 2026, "url": "https://arxiv.org/abs/2606.02646", "venue": "arXiv preprint", "categories": [ "Runtime", "Evaluation" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_chen_lemon", "title": "LEMON: Learning Executable Multi-Agent Orchestration via Counterfactual Reinforcement Learning", "authors": "Xudong Chen and Yixin Liu and Hua Wei and Kaize Ding", "year": 2026, "url": "https://arxiv.org/abs/2605.14483", "venue": "arXiv preprint", "categories": [ "Runtime", "Optimization" ], "subfamilies": [ "Optimization / Policy Learning", "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/4_3_optimization.tex" ], "method": "LEMON", "evidence": { "method": "LEMON", "family": "Optimization", "subfamily": "Policy", "update": "training", "effect": "orchestration", "boundary": "counterfactual cost", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2026_chen_token_economics", "title": "Token Economics for LLM Agents: A Dual-View Study from Computing and Economics", "authors": "Yuxi Chen and Junming Chen and Chenyu He and Yiwei Li and Yicheng Ji and Yifan Wu and Dingyu Yang and Lansong Diao and Lidan Shou and Hongliang Zhang and Huan Li and Gang Chen", "year": 2026, "url": "https://arxiv.org/abs/2605.09104", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_chi_evocf", "title": "EvoCF: Multi-Agent Collaboration via Agentic Memory-Driven Evolutionary Counterfactual Planning", "authors": "Haotian Chi and Zeyu Feng and Xingrui Yu and Linbo Luo and Yew-Soon Ong and Ivor Tsang and Hechang Chen and Yi Chang and Haiyan Yin", "year": 2026, "url": "https://openreview.net/forum?id=zGKkewtb2w", "venue": "International Conference on Learning Representations Workshop on Multi-Agent Learning and Its Opportunities in the Era of Generative AI (ICLR MALGAI-26)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Continual Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_chun_cara", "title": "What Do Agents Communicate? Characterizing Information Exchange in Multi-Agent Systems", "authors": "Yong Jin Chun and Iftekhar Ahmed", "year": 2026, "url": "https://arxiv.org/abs/2605.20548", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_dong_bandmas", "title": "BANDMAS: Causality-Inspired Semantic Packet Scheduling for Bandwidth-Efficient Multi-Agent Collaboration", "authors": "Jiangwen Dong and Wanyu Lin", "year": 2026, "url": "https://arxiv.org/abs/2608.00458", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_fan_todycomm", "title": "TodyComm: Task-Oriented Dynamic Communication for Multi-Round LLM-based Multi-Agent System", "authors": "Wenzhe Fan and Tommaso Tognoli and Henry Peng Zou and Chunyu Miao and Yibo Wang and Xinhua Zhang", "year": 2026, "url": "https://arxiv.org/abs/2602.03688", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_fu_latentmem", "title": "LatentMem: Customizing Latent Memory for Multi-Agent Systems", "authors": "Muxin Fu and Xiangyuan Xue and Yafu Li and Zefeng He and Siyuan Huang and Xiaoye Qu and Yu Cheng and Yang Yang", "year": 2026, "url": "https://arxiv.org/abs/2602.03036", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_gao_lmad", "title": "Where Reasoning Diverges: Localized Multi-Agent Debate for Multi-Hop Question Answering", "authors": "Weijun Gao and Xiang Ding and Haoyang Liu and Tiancheng Xing", "year": 2026, "url": "https://arxiv.org/abs/2608.01463", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_han_legomem", "title": "LEGOMem: Modular Procedural Memory for Multi-agent LLM Systems for Workflow Automation", "authors": "Dongge Han and Camille Couturier and Daniel Madrigal Diaz and Xuchao Zhang and Victor Rühle and Saravan Rajmohan", "year": 2025, "url": "https://arxiv.org/abs/2510.04851", "venue": "International Conference on Autonomous Agents and Multiagent Systems (AAMAS-26)", "categories": [ "Runtime", "Optimization" ], "subfamilies": [ "Optimization / Continual Learning", "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_hao_metateam", "title": "Evolve as a Team: Collaborative Self-Evolution for LLM-based Multi-Agent Systems", "authors": "Zhezheng Hao and Tianfu Wang and Huanshuo Dong and Ziyan Liu and Hong Wang and Xiankun Lin and Qiang Lin and Can Wang and Hande Dong and Jiawei Chen", "year": 2026, "url": "https://arxiv.org/abs/2605.29790", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Continual Learning" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex" ], "method": "MetaTeam", "evidence": { "method": "MetaTeam", "family": "Optimization", "subfamily": "Continual", "update": "continual", "effect": "reusable team updates", "boundary": "evolution cost not isolated", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2026_hao_not_all_flips", "title": "Not All Flips Are Conformity: Decomposing Stance Convergence in Multi-Agent LLM Debate", "authors": "Xiqi Hao and Zengqing Wu and Yu-Xuan Qiu and Chuan Xiao and Ruiqi Xu and Shuyuan Zheng and Jianbin Qin", "year": 2026, "url": "https://arxiv.org/abs/2606.00820", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_he_sigma", "title": "Conflict-Resilient Multi-Agent Reasoning via Signed Graph Modeling", "authors": "Longgang He and Longzhu He and Daojing He and Chaozhuo Li", "year": 2026, "url": "https://arxiv.org/abs/2605.19418", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_huang_evolverouter", "title": "EvolveRouter: Co-Evolving Routing and Prompt for Multi-Agent Question Answering", "authors": "Jiatan Huang and Zheyuan Zhang and Kaiwen Shi and Yanfang Ye and Chuxu Zhang", "year": 2026, "url": "https://arxiv.org/abs/2604.05149", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Policy Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_jeon_lragent", "title": "LRAgent: Efficient KV Cache Sharing for Multi-LoRA LLM Agents", "authors": "Hyesung Jeon and Hyeongju Ha and Jae-Joon Kim", "year": 2026, "url": "https://arxiv.org/abs/2602.01053", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_jiang_agentqmix", "title": "Agent Q-Mix: Selecting the Right Action for LLM Multi-Agent Systems through Reinforcement Learning", "authors": "Eric Hanchen Jiang and Levina Li and Rui Sun and Xiao Liang and Yubei Li and Yuchen Wu and Haozheng Luo and Hengli Li and Zhi Zhang and Zhaolu Kang and Kai-Wei Chang and Ying Nian Wu", "year": 2026, "url": "https://arxiv.org/abs/2604.00344", "venue": "arXiv preprint", "categories": [ "Runtime", "Optimization" ], "subfamilies": [ "Optimization / Policy Learning", "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_li_agent_harness_survey", "title": "Agent Harness Engineering: A Survey", "authors": "Junjie Li and Xi Xiao and Yunbei Zhang and Chen Liu and Lin Zhao and Xiaoying Liao and Yingrui Ji and Janet Wang and Jianyang Gu and Yingqiang Ge and Weijie Xu and Xi Fang and Xiang Xu and Tianchen Zhao and Youngeun Kim and Tianyang Wang and Jihun Hamm and Smita Krishnaswamy and Jun Huan and Chandan K. Reddy", "year": 2026, "url": "https://openreview.net/forum?id=eONq7FdiHa", "venue": "OpenReview preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_liu_acemad", "title": "Breaking the Martingale Curse: Multi-Agent Debate via Asymmetric Cognitive Potential Energy", "authors": "Yuhan Liu and Juntian Zhang and Yichen Wu and Martin Takac and Salem Lahlou and Xiuying Chen and Nils Lukas", "year": 2026, "url": "https://arxiv.org/abs/2603.06801", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_liu_droidspeak", "title": "DroidSpeak: KV Cache Sharing for Cross-LLM Communication and Multi-LLM Serving", "authors": "Yuhan Liu and Yuyang Huang and Jiayi Yao and Shaoting Feng and Zhuohan Gu and Kuntai Du and Hanchen Li and Yihua Cheng and Junchen Jiang and Shan Lu and Madan Musuvathi and Esha Choukse", "year": 2024, "url": "https://arxiv.org/abs/2411.02820", "venue": "USENIX Symposium on Networked Systems Design and Implementation (NSDI-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_mazumder_agentcollabbench", "title": "AgentCollabBench: Diagnosing When Good Agents Make Bad Collaborators", "authors": "Aritra Mazumder and Shubhashis Roy Dipta and Nusrat Jahan Lia and Tanzila Khan and Kainat Raisa Hossain and Nehaa Shri and Shubhrangshu Debsarkar and Humayra Tasnim and Gour Gupal Talukder Shawon and Debjoty Mitra and Sumaiya Ahmed Rani and Al Jami Islam Anik and Al Nafeu Khan", "year": 2026, "url": "https://arxiv.org/abs/2605.08647", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_meng_equimem", "title": "EquiMem: Calibrating Shared Memory in Multi-Agent Debate via Game-Theoretic Equilibrium", "authors": "Yuqiao Meng and Sakshi Sunil Narvekar and Luoxi Tang and Rupali Rajendra Vaje and Yingxue Zhang and Muchao Ye and Zhaohan Xi", "year": 2026, "url": "https://arxiv.org/abs/2605.09278", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_nie_skillgraph", "title": "SkillGraph: Self-Evolving Multi-Agent Collaboration with Multimodal Graph Topology", "authors": "Zheng Nie and Ruolin Shen and Xinlei Yu and Bo Yin and Jiangning Zhang and Xiaobin Hu", "year": 2026, "url": "https://arxiv.org/abs/2604.17503", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Continual Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_niu_seics", "title": "Reliability-Contagion Feasibility in LLM Multi-Agent Networks", "authors": "Ruiwu Niu and Xincheng Shu and Ying Zhao", "year": 2026, "url": "https://arxiv.org/abs/2607.21912", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_okawa_biased_consensus", "title": "Emergence of Biased Consensus in Multi-Agent LLM Debates", "authors": "Maya Okawa", "year": 2026, "url": "https://arxiv.org/abs/2608.02827", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_pan_task_structure_limits", "title": "How Task Structure Limits Multi-Agent Success: An Information-Theoretic Analysis", "authors": "Shi Pan and Ming Luo", "year": 2026, "url": "https://arxiv.org/abs/2606.13733", "venue": "arXiv preprint", "categories": [ "Runtime", "Collaboration Boundary" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/3_boundaries.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_parmar_debate_data_cleaning", "title": "When Helping Hurts and How to Fix It: Multi-Agent Debate for Data Cleaning", "authors": "Chirag Parmar and Akshat Mehta and Henglin Wu and Jagadish Ramamurthy and Shweta Medhekar", "year": 2026, "url": "https://arxiv.org/abs/2606.02866", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_salim_tokenomics", "title": "Tokenomics: Quantifying Where Tokens Are Used in Agentic Software Engineering", "authors": "Mohamad Salim and Jasmine Latendresse and SayedHassan Khatoonabadi and Emad Shihab", "year": 2026, "url": "https://arxiv.org/abs/2601.14470", "venue": "arXiv preprint", "categories": [ "Runtime", "Evaluation" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_sander_scaling_mas", "title": "Scaling LLM-Driven Multi-Agent Systems: Design Principles and Architectural Scalability Analysis", "authors": "Linus Sander and Fengjunjie Pan and Vahid Zolfaghari and Andre Schamschurko and Nenad Petrovic and Alois Knoll", "year": 2026, "url": "https://arxiv.org/abs/2607.27942", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_tang_hiva", "title": "HiVA: Self-Organized Hierarchical Variable Agent via Goal-Driven Semantic-Topological Evolution", "authors": "Jinzhou Tang and Jusheng Zhang and Qinhan Lv and Sidi Liu and Jing Yang and Chengpei Tang and Keze Wang", "year": 2026, "url": "https://ojs.aaai.org/index.php/AAAI/article/view/37190", "venue": "40th Annual AAAI Conference on Artificial Intelligence (AAAI-26)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "HiVA", "evidence": { "method": "HiVA", "family": "Topology", "subfamily": "Construction", "update": "evolution + query", "effect": "accuracy and dollar efficiency", "boundary": "semantics and topology co-evolve", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2026_vanmulders_digital_pantheon", "title": "Digital Pantheon: Simulating and Auditing Coalition Formation with LLM Agents", "authors": "Dylan Van Mulders and Matthias Bogaert and Dirk Van den Poel", "year": 2026, "url": "https://arxiv.org/abs/2607.15095", "venue": "AIDEM Workshop at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD AIDEM-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_wan_delibtrace", "title": "The Deliberative Illusion: Diagnosing Factual Attrition and Stance Homogenization in Multi-Agent LLM Deliberation", "authors": "Herun Wan and Jiaying Wu and Minnan Luo and Fanxiao Li and Ningnan Wang and Nancy F. Chen and Min-Yen Kan", "year": 2026, "url": "https://arxiv.org/abs/2606.03032", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_wang_agentops", "title": "Agent System Operations: Categorization, Challenges, and Future Directions", "authors": "Zexin Wang and Changhua Pei and Yuanhao Liu and Jingjing Li and Yintong Huo and Quan Zhou and Haotian Si and Hang Cui and Zihan Liu and Gaogang Xie and Fei Sun and Dan Pei and David Lo", "year": 2026, "url": "https://arxiv.org/abs/2606.01581", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_wang_budgeted_actordefer", "title": "Budgeted Act-or-Defer Multi-Agent LLM Deliberation with Local Reliability Bounds", "authors": "Mengdie Flora Wang and Haochen Xie and Guanghui Wang and Devin Zhang and Jae Oh Woo", "year": 2026, "url": "https://arxiv.org/abs/2606.29654", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex" ], "method": "Act-or-Defer", "evidence": { "method": "Act-or-Defer", "family": "Runtime", "subfamily": "Scheduling", "update": "runtime", "effect": "calls and risk", "boundary": "calibration required", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2026_wang_handraiser", "title": "Learning to Interrupt in Language-based Multi-agent Communication", "authors": "Danqing Wang and Da Yin and Ruta Desai and Lei Li and Asli Celikyilmaz and Ansong Ni", "year": 2026, "url": "https://arxiv.org/abs/2604.06452", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_wang_metagen", "title": "MetaGen: Self-Evolving Roles and Topologies for Multi-Agent LLM Reasoning", "authors": "Yimeng Wang and Jiaxing Zhao and Hongbin Xie and Hexing Ma and Yuzhen Lei and Shuangxue Liu and Xuan Song and Zichen Zhang and Haoran Zhang", "year": 2026, "url": "https://arxiv.org/abs/2601.19290", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Adaptation" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "MetaGen", "evidence": { "method": "MetaGen", "family": "Topology", "subfamily": "Adaptation", "update": "continual + runtime", "effect": "roles and topology", "boundary": "roles and topology co-adapt", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2026_wu_debateocr", "title": "Cross-Modal Memory Compression for Efficient Multi-Agent Debate", "authors": "Jing Wu and Yue Sun and Tianpei Xie and Suiyao Chen and Jingyuan Bao and Yaopengxiao Xu and Gaoyuan Du and Inseok Heo and Alexander Gutfraind and Xin Wang", "year": 2026, "url": "https://arxiv.org/abs/2602.00454", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex" ], "method": "DebateOCR", "evidence": { "method": "DebateOCR", "family": "Runtime", "subfamily": "Communication", "update": "runtime", "effect": "input tokens, latency", "boundary": "multimodal channel", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "2026_xu_oneflow", "title": "Rethinking the Value of Multi-Agent Workflow: A Strong Single Agent Baseline", "authors": "Jiawei Xu and Arief Koesdwiady and Sisong Bei and Yan Han and Baixiang Huang and Dakuo Wang and Yutong Chen and Zheshen Wang and Peihao Wang and Pan Li and Ying Ding", "year": 2026, "url": "https://arxiv.org/abs/2601.12307", "venue": "arXiv preprint", "categories": [ "Runtime", "Evaluation", "Synthesis" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex", "sections/6_synthesis.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_yan_hiagreement", "title": "When Truth Is Distributed: Misinformation Derails Collective Fact Recovery in LLM-Based Multi-Agent Systems", "authors": "Chenfei Yan and Zeyang Yue and Feifei Zhao and Erliang Lin and Lu Jia and Haibo Tong and Mingyang Lyu and Chengyi Sun and Yi Zeng", "year": 2026, "url": "https://arxiv.org/abs/2608.03421", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_yang_hila", "title": "Adaptive Collaboration with Humans: Metacognitive Policy Optimization for Multi-Agent LLMs with Continual Learning", "authors": "Wei Yang and Defu Cao and Jiacheng Pang and Muyan Weng and Yan Liu", "year": 2026, "url": "https://arxiv.org/abs/2603.07972", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Continual Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_yu_multiagent_memory_arch", "title": "Multi-Agent Memory from a Computer Architecture Perspective: Visions and Challenges Ahead", "authors": "Zhongming Yu and Naicheng Yu and Hejia Zhang and Wentao Ni and Mingrui Yin and Jiaying Yang and Yujie Zhao and Jishen Zhao", "year": 2026, "url": "https://arxiv.org/abs/2603.10062", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_yu_when_mas_help", "title": "When Do Multi-Agent Systems Help? An Information Bottleneck Perspective", "authors": "Wendi Yu and Lianhao Zhou and Xiangjue Dong and Sai Sudarshan Barath and Declan Staunton and Byung-Jun Yoon and Xiaoning Qian and James Caverlee and Shuiwang Ji", "year": 2026, "url": "https://arxiv.org/abs/2607.16133", "venue": "arXiv preprint", "categories": [ "Runtime", "Collaboration Boundary", "Evaluation" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/3_boundaries.tex", "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_zhang_agentrevive", "title": "Taming ``Zombie'' Agents: A Markov State-Aware Framework for Resilient Multi-Agent Evolution", "authors": "Taolin Zhang and Pukun Zhao and Qizhou Chen and Jiuheng Wan and Chen Chen and Xiaofeng He and Chengyu Wang and Richang Hong", "year": 2026, "url": "https://arxiv.org/abs/2605.17348", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-26)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Scheduling" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "2026_zhang_radar", "title": "RADAR: Redundancy-Aware Diffusion for Multi-Agent Communication Structure Generation", "authors": "Zhen Zhang and Wanjing Zhou and Juncheng Li and Hao Fei and Jun Wen and Wei Ji", "year": 2026, "url": "https://arxiv.org/abs/2605.09907", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv220311171", "title": "Self-Consistency Improves Chain of Thought Reasoning in Language Models", "authors": "Xuezhi Wang and Jason Wei and Dale Schuurmans and Quoc Le and Ed Chi and Sharan Narang and Aakanksha Chowdhery and Denny Zhou", "year": 2022, "url": "https://arxiv.org/abs/2203.11171", "venue": "International Conference on Learning Representations (ICLR-23)", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv221003629", "title": "ReAct: Synergizing Reasoning and Acting in Language Models", "authors": "Shunyu Yao and Jeffrey Zhao and Dian Yu and Nan Du and Izhak Shafran and Karthik Narasimhan and Yuan Cao", "year": 2022, "url": "https://arxiv.org/abs/2210.03629", "venue": "International Conference on Learning Representations (ICLR-23)", "categories": [ "Background", "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/2_scope.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230311366", "title": "Reflexion: Language Agents with Verbal Reinforcement Learning", "authors": "Noah Shinn and Federico Cassano and Edward Berman and Ashwin Gopinath and Karthik Narasimhan and Shunyu Yao", "year": 2023, "url": "https://arxiv.org/abs/2303.11366", "venue": "Conference on Neural Information Processing Systems (NeurIPS-23)", "categories": [ "Background", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230317651", "title": "Self-Refine: Iterative Refinement with Self-Feedback", "authors": "Aman Madaan and Niket Tandon and Prakhar Gupta and Skyler Hallinan and Luyu Gao and Sarah Wiegreffe and Uri Alon and Nouha Dziri and Shrimai Prabhumoye and Yiming Yang and Shashank Gupta and Bodhisattwa Prasad Majumder and Katherine Hermann and Sean Welleck and Amir Yazdanbakhsh and Peter Clark", "year": 2023, "url": "https://arxiv.org/abs/2303.17651", "venue": "Conference on Neural Information Processing Systems (NeurIPS-23)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230317760", "title": "CAMEL: Communicative Agents for \"Mind\" Exploration of Large Language Model Society", "authors": "Guohao Li and Hasan Abed Al Kader Hammoud and Hani Itani and Dmitrii Khizbullin and Bernard Ghanem", "year": 2023, "url": "https://arxiv.org/abs/2303.17760", "venue": "Conference on Neural Information Processing Systems (NeurIPS-23)", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230403442", "title": "Generative Agents: Interactive Simulacra of Human Behavior", "authors": "Joon Sung Park and Joseph C. O'Brien and Carrie J. Cai and Meredith Ringel Morris and Percy Liang and Michael S. Bernstein", "year": 2023, "url": "https://arxiv.org/abs/2304.03442", "venue": "ACM Symposium on User Interface Software and Technology (UIST-23)", "categories": [ "Background", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230408244", "title": "API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMs", "authors": "Minghao Li and Yingxiu Zhao and Bowen Yu and Feifan Song and Hangyu Li and Haiyang Yu and Zhoujun Li and Fei Huang and Yongbin Li", "year": 2023, "url": "https://arxiv.org/abs/2304.08244", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-23)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230505176", "title": "FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance", "authors": "Lingjiao Chen and Matei Zaharia and James Zou", "year": 2023, "url": "https://arxiv.org/abs/2305.05176", "venue": "Transactions on Machine Learning Research (TMLR)", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230514325", "title": "Improving Factuality and Reasoning in Language Models through Multiagent Debate", "authors": "Yilun Du and Shuang Li and Antonio Torralba and Joshua B. Tenenbaum and Igor Mordatch", "year": 2023, "url": "https://arxiv.org/abs/2305.14325", "venue": "International Conference on Machine Learning (ICML-24)", "categories": [ "Topology", "Runtime", "Background", "Introduction" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/1_intro.tex", "sections/2_scope.tex", "sections/4_1_topology.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230515334", "title": "Gorilla: Large Language Model Connected with Massive APIs", "authors": "Shishir G. Patil and Tianjun Zhang and Xin Wang and Joseph E. Gonzalez", "year": 2023, "url": "https://arxiv.org/abs/2305.15334", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230516291", "title": "Voyager: An Open-Ended Embodied Agent with Large Language Models", "authors": "Guanzhi Wang and Yuqi Xie and Yunfan Jiang and Ajay Mandlekar and Chaowei Xiao and Yuke Zhu and Linxi Fan and Anima Anandkumar", "year": 2023, "url": "https://arxiv.org/abs/2305.16291", "venue": "Transactions on Machine Learning Research (TMLR)", "categories": [ "Background", "Evaluation", "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/2_scope.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230606070", "title": "Mind2Web: Towards a Generalist Agent for the Web", "authors": "Xiang Deng and Yu Gu and Boyuan Zheng and Shijie Chen and Samuel Stevens and Boshi Wang and Huan Sun and Yu Su", "year": 2023, "url": "https://arxiv.org/abs/2306.06070", "venue": "Conference on Neural Information Processing Systems (NeurIPS-23)", "categories": [ "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230606624", "title": "RestGPT: Connecting Large Language Models with Real-World RESTful APIs", "authors": "Yifan Song and Weimin Xiong and Dawei Zhu and Wenhao Wu and Han Qian and Mingbo Song and Hailiang Huang and Cheng Li and Ke Wang and Rong Yao and Ye Tian and Sujian Li", "year": 2023, "url": "https://arxiv.org/abs/2306.06624", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230702485", "title": "Building Cooperative Embodied Agents Modularly with Large Language Models", "authors": "Hongxin Zhang and Weihua Du and Jiaming Shan and Qinhong Zhou and Yilun Du and Joshua B. Tenenbaum and Tianmin Shu and Chuang Gan", "year": 2023, "url": "https://arxiv.org/abs/2307.02485", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Evaluation", "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230704738", "title": "RoCo: Dialectic Multi-Robot Collaboration with Large Language Models", "authors": "Zhao Mandi and Shreeya Jain and Shuran Song", "year": 2023, "url": "https://arxiv.org/abs/2307.04738", "venue": "IEEE International Conference on Robotics and Automation (ICRA-24)", "categories": [ "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230707924", "title": "ChatDev: Communicative Agents for Software Development", "authors": "Chen Qian and Wei Liu and Hongzhang Liu and Nuo Chen and Yufan Dang and Jiahao Li and Cheng Yang and Weize Chen and Yusheng Su and Xin Cong and Juyuan Xu and Dahai Li and Zhiyuan Liu and Maosong Sun", "year": 2023, "url": "https://arxiv.org/abs/2307.07924", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230713854", "title": "WebArena: A Realistic Web Environment for Building Autonomous Agents", "authors": "Shuyan Zhou and Frank F. Xu and Hao Zhu and Xuhui Zhou and Robert Lo and Abishek Sridhar and Xianyi Cheng and Tianyue Ou and Yonatan Bisk and Daniel Fried and Uri Alon and Graham Neubig", "year": 2023, "url": "https://arxiv.org/abs/2307.13854", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Runtime", "Evaluation", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230716789", "title": "ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs", "authors": "Yujia Qin and Shihao Liang and Yining Ye and Kunlun Zhu and Lan Yan and Yaxi Lu and Yankai Lin and Xin Cong and Xiangru Tang and Bill Qian and Sihan Zhao and Lauren Hong and Runchu Tian and Ruobing Xie and Jie Zhou and Mark Gerstein and Dahai Li and Zhiyuan Liu and Maosong Sun", "year": 2023, "url": "https://arxiv.org/abs/2307.16789", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Runtime", "Background", "Introduction" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/1_intro.tex", "sections/2_scope.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230800352", "title": "MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework", "authors": "Sirui Hong and Mingchen Zhuge and Jiaqi Chen and Xiawu Zheng and Yuheng Cheng and Ceyao Zhang and Jinlin Wang and Zili Wang and Steven Ka Shing Yau and Zijuan Lin and Liyang Zhou and Chenyu Ran and Lingfeng Xiao and Chenglin Wu and Jürgen Schmidhuber", "year": 2023, "url": "https://arxiv.org/abs/2308.00352", "venue": "arXiv preprint", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230803688", "title": "AgentBench: Evaluating LLMs as Agents", "authors": "Xiao Liu and Hao Yu and Hanchen Zhang and Yifan Xu and Xuanyu Lei and Hanyu Lai and Yu Gu and Hangliang Ding and Kaiwen Men and Kejuan Yang and Shudan Zhang and Xiang Deng and Aohan Zeng and Zhengxiao Du and Chenhui Zhang and Sheng Shen and Tianjun Zhang and Yu Su and Huan Sun and Minlie Huang and Yuxiao Dong and Jie Tang", "year": 2023, "url": "https://arxiv.org/abs/2308.03688", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Evaluation", "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230807201", "title": "ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debate", "authors": "Chi-Min Chan and Weize Chen and Yusheng Su and Jianxuan Yu and Wei Xue and Shanghang Zhang and Jie Fu and Zhiyuan Liu", "year": 2023, "url": "https://arxiv.org/abs/2308.07201", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230808155", "title": "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation", "authors": "Qingyun Wu and Gagan Bansal and Jieyu Zhang and Yiran Wu and Beibin Li and Erkang Zhu and Li Jiang and Xiaoyun Zhang and Shaokun Zhang and Jiale Liu and Ahmed Hassan Awadallah and Ryen W White and Doug Burger and Chi Wang", "year": 2023, "url": "https://arxiv.org/abs/2308.08155", "venue": "arXiv preprint", "categories": [ "Background" ], "subfamilies": [], "sourceFiles": [ "sections/2_scope.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230811339", "title": "ProAgent: Building Proactive Cooperative Agents with Large Language Models", "authors": "Ceyao Zhang and Kaijie Yang and Siyi Hu and Zihao Wang and Guanghe Li and Yihang Sun and Cheng Zhang and Zhaowei Zhang and Anji Liu and Song-Chun Zhu and Xiaojun Chang and Junge Zhang and Feng Yin and Yitao Liang and Yaodong Yang", "year": 2023, "url": "https://arxiv.org/abs/2308.11339", "venue": "38th Annual AAAI Conference on Artificial Intelligence (AAAI-24)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv230913007", "title": "ReConcile: Round-Table Conference Improves Reasoning via Consensus among Diverse LLMs", "authors": "Justin Chih-Yao Chen and Swarnadeep Saha and Mohit Bansal", "year": 2023, "url": "https://arxiv.org/abs/2309.13007", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv231002170", "title": "A Dynamic LLM-Powered Agent Network for Task-Oriented Agent Collaboration", "authors": "Zijun Liu and Yanzhe Zhang and Peng Li and Yang Liu and Diyi Yang", "year": 2023, "url": "https://arxiv.org/abs/2310.02170", "venue": "Conference on Language Modeling (COLM-24)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Adaptation" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv231011667", "title": "SOTOPIA: Interactive Evaluation for Social Intelligence in Language Agents", "authors": "Xuhui Zhou and Hao Zhu and Leena Mathur and Ruohong Zhang and Haofei Yu and Zhengyang Qi and Louis-Philippe Morency and Yonatan Bisk and Daniel Fried and Graham Neubig and Maarten Sap", "year": 2023, "url": "https://arxiv.org/abs/2310.11667", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Evaluation", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv231112983", "title": "GAIA: a benchmark for General AI Assistants", "authors": "Grégoire Mialon and Clémentine Fourrier and Craig Swift and Thomas Wolf and Yann LeCun and Thomas Scialom", "year": 2023, "url": "https://arxiv.org/abs/2311.12983", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv231213010", "title": "AgentCoder: Multi-Agent-based Code Generation with Iterative Testing and Optimisation", "authors": "Dong Huang and Jie M. Zhang and Michael Luck and Qingwen Bu and Yuhao Qing and Heming Cui", "year": 2023, "url": "https://arxiv.org/abs/2312.13010", "venue": "arXiv preprint", "categories": [ "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240113178", "title": "AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents", "authors": "Chang Ma and Junlei Zhang and Zhihao Zhu and Cheng Yang and Yujiu Yang and Yaohui Jin and Zhenzhong Lan and Lingpeng Kong and Junxian He", "year": 2024, "url": "https://arxiv.org/abs/2401.13178", "venue": "Conference on Neural Information Processing Systems (NeurIPS-24)", "categories": [ "Evaluation", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240205120", "title": "More Agents Is All You Need", "authors": "Junyou Li and Qin Zhang and Yangbin Yu and Qiang Fu and Deheng Ye", "year": 2024, "url": "https://arxiv.org/abs/2402.05120", "venue": "Transactions on Machine Learning Research (TMLR)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240206664", "title": "LLM Agents can Autonomously Hack Websites", "authors": "Richard Fang and Rohan Bindu and Akul Gupta and Qiusi Zhan and Daniel Kang", "year": 2024, "url": "https://arxiv.org/abs/2402.06664", "venue": "arXiv preprint", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240211550", "title": "LongAgent: Scaling Language Models to 128k Context through Multi-Agent Collaboration", "authors": "Jun Zhao and Can Zu and Hao Xu and Yi Lu and Wei He and Yiwen Ding and Tao Gui and Qi Zhang and Xuanjing Huang", "year": 2024, "url": "https://arxiv.org/abs/2402.11550", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240216823", "title": "Language Agents as Optimizable Graphs", "authors": "Mingchen Zhuge and Wenyi Wang and Louis Kirsch and Francesco Faccio and Dmitrii Khizbullin and Jürgen Schmidhuber", "year": 2024, "url": "https://arxiv.org/abs/2402.16823", "venue": "International Conference on Machine Learning (ICML-24)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240302691", "title": "InjecAgent: Benchmarking Indirect Prompt Injections in Tool-Integrated Large Language Model Agents", "authors": "Qiusi Zhan and Zhixiang Liang and Zifan Ying and Daniel Kang", "year": 2024, "url": "https://arxiv.org/abs/2403.02691", "venue": "Findings of the Association for Computational Linguistics: ACL 2024 (ACL Findings-24)", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240307718", "title": "WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?", "authors": "Alexandre Drouin and Maxime Gasse and Massimo Caccia and Issam H. Laradji and Manuel Del Verme and Tom Marty and Léo Boisvert and Megh Thakkar and Quentin Cappart and David Vazquez and Nicolas Chapados and Alexandre Lacoste", "year": 2024, "url": "https://arxiv.org/abs/2403.07718", "venue": "International Conference on Machine Learning (ICML-24)", "categories": [ "Runtime", "Evaluation", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240407972", "title": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments", "authors": "Tianbao Xie and Danyang Zhang and Jixuan Chen and Xiaochuan Li and Siheng Zhao and Ruisheng Cao and Toh Jing Hua and Zhoujun Cheng and Dongchan Shin and Fangyu Lei and Yitao Liu and Yiheng Xu and Shuyan Zhou and Silvio Savarese and Caiming Xiong and Victor Zhong and Tao Yu", "year": 2024, "url": "https://arxiv.org/abs/2404.07972", "venue": "Conference on Neural Information Processing Systems (NeurIPS-24)", "categories": [ "Runtime", "Evaluation", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240408144", "title": "LLM Agents can Autonomously Exploit One-day Vulnerabilities", "authors": "Richard Fang and Rohan Bindu and Akul Gupta and Daniel Kang", "year": 2024, "url": "https://arxiv.org/abs/2404.08144", "venue": "arXiv preprint", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240514314", "title": "Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration", "authors": "Yang Zhang and Shixin Yang and Chenjia Bai and Fei Wu and Xiu Li and Zhen Wang and Xuelong Li", "year": 2024, "url": "https://arxiv.org/abs/2405.14314", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-25)", "categories": [ "Optimization", "Open Problems" ], "subfamilies": [ "Optimization / Policy Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240515793", "title": "SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering", "authors": "John Yang and Carlos E. Jimenez and Alexander Wettig and Kilian Lieret and Shunyu Yao and Karthik Narasimhan and Ofir Press", "year": 2024, "url": "https://arxiv.org/abs/2405.15793", "venue": "Conference on Neural Information Processing Systems (NeurIPS-24)", "categories": [ "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240601637", "title": "Teams of LLM Agents can Exploit Zero-Day Vulnerabilities", "authors": "Yuxuan Zhu and Antony Kellermann and Akul Gupta and Philip Li and Richard Fang and Rohan Bindu and Daniel Kang", "year": 2024, "url": "https://arxiv.org/abs/2406.01637", "venue": "Conference of the European Chapter of the Association for Computational Linguistics (EACL-26)", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240602818", "title": "Chain of Agents: Large Language Models Collaborating on Long-Context Tasks", "authors": "Yusen Zhang and Ruoxi Sun and Yanfei Chen and Tomas Pfister and Rui Zhang and Sercan Ö. Arik", "year": 2024, "url": "https://arxiv.org/abs/2406.02818", "venue": "Conference on Neural Information Processing Systems (NeurIPS-24)", "categories": [ "Runtime", "Collaboration Boundary", "Introduction" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/1_intro.tex", "sections/3_boundaries.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240604151", "title": "AgentGym: Evolving Large Language Model-based Agents across Diverse Environments", "authors": "Zhiheng Xi and Yiwen Ding and Wenxiang Chen and Boyang Hong and Honglin Guo and Junzhe Wang and Dingwen Yang and Chenyang Liao and Xin Guo and Wei He and Songyang Gao and Lu Chen and Rui Zheng and Yicheng Zou and Tao Gui and Qi Zhang and Xipeng Qiu and Xuanjing Huang and Zuxuan Wu and Yu-Gang Jiang", "year": 2024, "url": "https://arxiv.org/abs/2406.04151", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240604692", "title": "Mixture-of-Agents Enhances Large Language Model Capabilities", "authors": "Junlin Wang and Jue Wang and Ben Athiwaratkun and Ce Zhang and James Zou", "year": 2024, "url": "https://arxiv.org/abs/2406.04692", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240607155", "title": "Scaling Large Language Model-based Multi-Agent Collaboration", "authors": "Chen Qian and Zihao Xie and YiFei Wang and Wei Liu and Kunlun Zhu and Hanchen Xia and Yufan Dang and Zhuoyun Du and Weize Chen and Cheng Yang and Zhiyuan Liu and Maosong Sun", "year": 2024, "url": "https://arxiv.org/abs/2406.07155", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "MACNET", "evidence": { "method": "MACNET", "family": "Topology", "subfamily": "Construction", "update": "design", "effect": "quality across agent scale", "boundary": "scale and topology co-vary", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240611776", "title": "Improving Multi-Agent Debate with Sparse Communication Topology", "authors": "Yunxuan Li and Yibing Du and Jiageng Zhang and Le Hou and Peter Grabowski and Yeqing Li and Eugene Ie", "year": 2024, "url": "https://arxiv.org/abs/2406.11776", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-24)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240612045", "title": "τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "authors": "Shunyu Yao and Noah Shinn and Pedram Razavi and Karthik Narasimhan", "year": 2024, "url": "https://arxiv.org/abs/2406.12045", "venue": "arXiv preprint", "categories": [ "Runtime", "Evaluation" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240613352", "title": "AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents", "authors": "Edoardo Debenedetti and Jie Zhang and Mislav Balunović and Luca Beurer-Kellner and Marc Fischer and Florian Tramèr", "year": 2024, "url": "https://arxiv.org/abs/2406.13352", "venue": "Conference on Neural Information Processing Systems (NeurIPS-24)", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240614550", "title": "GraphReader: Building Graph-based Agent to Enhance Long-Context Abilities of Large Language Models", "authors": "Shilong Li and Yancheng He and Hangyu Guo and Xingyuan Bu and Ge Bai and Jie Liu and Jiaheng Liu and Xingwei Qu and Yangguang Li and Wanli Ouyang and Wenbo Su and Bo Zheng", "year": 2024, "url": "https://arxiv.org/abs/2406.14550", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-24)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240618665", "title": "RouteLLM: Learning to Route LLMs with Preference Data", "authors": "Isaac Ong and Amjad Almahairi and Vincent Wu and Wei-Lin Chiang and Tianhao Wu and Joseph E. Gonzalez and M Waleed Kadous and Ion Stoica", "year": 2024, "url": "https://arxiv.org/abs/2406.18665", "venue": "arXiv preprint", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240701476", "title": "Tree Search for Language Model Agents", "authors": "Jing Yu Koh and Stephen McAleer and Daniel Fried and Ruslan Salakhutdinov", "year": 2024, "url": "https://arxiv.org/abs/2407.01476", "venue": "Transactions on Machine Learning Research (TMLR)", "categories": [ "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240815971", "title": "BattleAgentBench: A Benchmark for Evaluating Cooperation and Competition Capabilities of Language Models in Multi-Agent Systems", "authors": "Wei Wang and Dan Zhang and Tao Feng and Boyan Wang and Jie Tang", "year": 2024, "url": "https://arxiv.org/abs/2408.15971", "venue": "arXiv preprint", "categories": [ "Evaluation", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv240918028", "title": "Compositional Hardness of Code in Large Language Models – A Probabilistic Perspective", "authors": "Yotam Wolf and Binyamin Rothberg and Dorin Shteyman and Amnon Shashua", "year": 2024, "url": "https://arxiv.org/abs/2409.18028", "venue": "arXiv preprint", "categories": [ "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv241007283", "title": "Prompt Infection: LLM-to-LLM Prompt Injection within Multi-Agent Systems", "authors": "Donghyun Lee and Mo Tiwari", "year": 2024, "url": "https://arxiv.org/abs/2410.07283", "venue": "arXiv preprint", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv241010762", "title": "AFlow: Automating Agentic Workflow Generation", "authors": "Jiayi Zhang and Jinyu Xiang and Zhaoyang Yu and Fengwei Teng and Xionghui Chen and Jiaqi Chen and Mingchen Zhuge and Xin Cheng and Sirui Hong and Jinlin Wang and Bingnan Zheng and Bang Liu and Yuyu Luo and Chenglin Wu", "year": 2024, "url": "https://arxiv.org/abs/2410.10762", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex" ], "method": "AFlow", "evidence": { "method": "AFlow", "family": "Optimization", "subfamily": "Workflow", "update": "design", "effect": "executable workflow", "boundary": "search amortization", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv241011782", "title": "G-Designer: Architecting Multi-agent Communication Topologies via Graph Neural Networks", "authors": "Guibin Zhang and Yanwei Yue and Xiangguo Sun and Guancheng Wan and Miao Yu and Junfeng Fang and Kun Wang and Tianlong Chen and Dawei Cheng", "year": 2024, "url": "https://arxiv.org/abs/2410.11782", "venue": "International Conference on Machine Learning (ICML-25)", "categories": [ "Topology", "Synthesis" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/6_synthesis.tex" ], "method": "G-Designer", "evidence": { "method": "G-Designer", "family": "Topology", "subfamily": "Construction", "update": "offline + query", "effect": "prompt tokens", "boundary": "fixed team and anchor", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv241016946", "title": "Self-Evolving Multi-Agent Collaboration Networks for Software Development", "authors": "Yue Hu and Yuzhu Cai and Yaxin Du and Xinyu Zhu and Xiangrui Liu and Zijie Yu and Yuchen Hou and Shuo Tang and Siheng Chen", "year": 2024, "url": "https://arxiv.org/abs/2410.16946", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250202533", "title": "Multi-Agent Design: Optimizing Agents with Better Prompts and Topologies", "authors": "Han Zhou and Xingchen Wan and Ruoxi Sun and Hamid Palangi and Shariq Iqbal and Ivan Vulić and Anna Korhonen and Sercan Ö. Arık", "year": 2025, "url": "https://arxiv.org/abs/2502.02533", "venue": "International Conference on Learning Representations (ICLR-26)", "categories": [ "Optimization", "Synthesis", "Open Problems" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex", "sections/6_synthesis.tex", "sections/7_outlook.tex" ], "method": "MASS", "evidence": { "method": "MASS", "family": "Optimization", "subfamily": "Workflow", "update": "design", "effect": "reusable workflow", "boundary": "coupled search stages", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250204180", "title": "Multi-agent Architecture Search via Agentic Supernet", "authors": "Guibin Zhang and Luyang Niu and Junfeng Fang and Kun Wang and Lei Bai and Xiang Wang", "year": 2025, "url": "https://arxiv.org/abs/2502.04180", "venue": "International Conference on Machine Learning (ICML-25)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250207373", "title": "EvoFlow: Evolving Diverse Agentic Workflows On The Fly", "authors": "Guibin Zhang and Kaijie Chen and Guancheng Wan and Heng Chang and Hong Cheng and Kun Wang and Shuyue Hu and Lei Bai", "year": 2025, "url": "https://arxiv.org/abs/2502.07373", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250211133", "title": "MasRouter: Learning to Route LLMs for Multi-Agent Systems", "authors": "Yanwei Yue and Guibin Zhang and Boyang Liu and Guancheng Wan and Kun Wang and Dawei Cheng and Yiyan Qi", "year": 2025, "url": "https://arxiv.org/abs/2502.11133", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-25)", "categories": [ "Runtime", "Collaboration Boundary" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/3_boundaries.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250219130", "title": "Voting or Consensus? Decision-Making in Multi-Agent Debate", "authors": "Lars Benedikt Kaesberg and Jonas Becker and Jan Philip Wahle and Terry Ruas and Bela Gipp", "year": 2025, "url": "https://arxiv.org/abs/2502.19130", "venue": "Findings of the Association for Computational Linguistics: ACL 2025 (ACL Findings-25)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250301935", "title": "MultiAgentBench: Evaluating the Collaboration and Competition of LLM agents", "authors": "Kunlun Zhu and Hongyi Du and Zhaochen Hong and Xiaocheng Yang and Shuyi Guo and Zhe Wang and Zhenhailong Wang and Cheng Qian and Xiangru Tang and Heng Ji and Jiaxuan You", "year": 2025, "url": "https://arxiv.org/abs/2503.01935", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-25)", "categories": [ "Evaluation", "Open Problems", "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250318891", "title": "AgentDropout: Dynamic Agent Elimination for Token-Efficient and High-Performance LLM-Based Multi-Agent Collaboration", "authors": "Zhexuan Wang and Yutong Wang and Xuebo Liu and Liang Ding and Miao Zhang and Jie Liu and Min Zhang", "year": 2025, "url": "https://arxiv.org/abs/2503.18891", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "AgentDropout", "evidence": { "method": "AgentDropout", "family": "Topology", "subfamily": "Pruning", "update": "task/round", "effect": "prompt and output tokens", "boundary": "not query-conditioned", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250511765", "title": "OMAC: A Holistic Optimization Framework for LLM-Based Multi-Agent Collaboration", "authors": "Shijun Li and Hilaf Hasson and Joydeep Ghosh", "year": 2025, "url": "https://arxiv.org/abs/2505.11765", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Workflow Search" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250518286", "title": "Single-agent or Multi-agent Systems? Why Not Both?", "authors": "Mingyan Gao and Yanzi Li and Banruo Liu and Yifan Yu and Phillip Wang and Ching-Yu Lin and Fan Lai", "year": 2025, "url": "https://arxiv.org/abs/2505.18286", "venue": "arXiv preprint", "categories": [ "Collaboration Boundary", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/3_boundaries.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250523352", "title": "Understanding the Information Propagation Effects of Communication Topologies in LLM-based Multi-Agent Systems", "authors": "Xu Shen and Yixin Liu and Yiwei Dai and Yili Wang and Rui Miao and Yue Tan and Shirui Pan and Xin Wang", "year": 2025, "url": "https://arxiv.org/abs/2505.23352", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-25)", "categories": [ "Topology", "Evaluation", "Open Problems" ], "subfamilies": [ "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250602951", "title": "Adaptive Graph Pruning for Multi-Agent Communication", "authors": "Boyi Li and Zhonghan Zhao and Der-Horng Lee and Gaoang Wang", "year": 2025, "url": "https://arxiv.org/abs/2506.02951", "venue": "European Conference on Artificial Intelligence (ECAI-25)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250617784", "title": "AnyMAC: Cascading Flexible Multi-Agent Collaboration via Next-Agent Prediction", "authors": "Song Wang and Zhen Tan and Zihan Chen and Shuang Zhou and Tianlong Chen and Jundong Li", "year": 2025, "url": "https://arxiv.org/abs/2506.17784", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-25)", "categories": [ "Runtime", "Collaboration Boundary", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/3_boundaries.tex", "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250713190", "title": "GEMMAS: Graph-based Evaluation Metrics for Multi Agent Systems", "authors": "Jisoo Lee and Raeyoung Chang and Dongwook Kwon and Harmanpreet Singh and Nikhil Verma", "year": 2025, "url": "https://arxiv.org/abs/2507.13190", "venue": "arXiv preprint", "categories": [ "Topology", "Evaluation", "Open Problems" ], "subfamilies": [ "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250718224", "title": "Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation", "authors": "Shiyuan Li and Yixin Liu and Qingsong Wen and Chengqi Zhang and Shirui Pan", "year": 2025, "url": "https://arxiv.org/abs/2507.18224", "venue": "40th Annual AAAI Conference on Artificial Intelligence (AAAI-26)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250804652", "title": "LLM Collaboration With Multi-Agent Reinforcement Learning", "authors": "Shuo Liu and Tianle Chen and Zeyu Liang and Xueguang Lyu and Christopher Amato", "year": 2025, "url": "https://arxiv.org/abs/2508.04652", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Policy Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250804903", "title": "RCR-Router: Efficient Role-Aware Context Routing for Multi-Agent LLM Systems with Structured Memory", "authors": "Jun Liu and Zhenglun Kong and Changdi Yang and Fan Yang and Tianqi Li and Peiyan Dong and Joannah Nanjekye and Hao Tang and Geng Yuan and Wei Niu and Wenbin Zhang and Pu Zhao and Xue Lin and Dong Huang and Yanzhi Wang", "year": 2025, "url": "https://arxiv.org/abs/2508.04903", "venue": "arXiv preprint", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "RCR-Router", "evidence": { "method": "RCR-Router", "family": "Runtime", "subfamily": "Routing", "update": "step", "effect": "exposed context", "boundary": "memory construction cost", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250811733", "title": "SafeSieve: From Heuristics to Experience in Progressive Pruning for LLM-based Multi-Agent Communication", "authors": "Ruijia Zhang and Xinyan Zhao and Ruixiang Wang and Sigen Chen and Guibin Zhang and An Zhang and Kun Wang and Qingsong Wen", "year": 2025, "url": "https://arxiv.org/abs/2508.11733", "venue": "40th Annual AAAI Conference on Artificial Intelligence (AAAI-26)", "categories": [ "Topology", "Runtime" ], "subfamilies": [ "Runtime / Communication", "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250906436", "title": "Tree of Agents: Improving Long-Context Capabilities of Large Language Models through Multi-Perspective Reasoning", "authors": "Song Yu and Xiaofei Xu and Ke Deng and Li Li and Lin Tian", "year": 2025, "url": "https://arxiv.org/abs/2509.06436", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-25)", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250907571", "title": "Towards Generalized Routing: Model and Agent Orchestration for Adaptive and Efficient Inference", "authors": "Xiyu Guo and Shan Wang and Chunfang Ji and Xuefeng Zhao and Wenhao Xi and Yaoyao Liu and Qinglan Li and Chao Deng and Junlan Feng", "year": 2025, "url": "https://arxiv.org/abs/2509.07571", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250911656", "title": "MALLM: Multi-Agent Large Language Models Framework", "authors": "Jonas Becker and Lars Benedikt Kaesberg and Niklas Bauer and Jan Philip Wahle and Terry Ruas and Bela Gipp", "year": 2025, "url": "https://arxiv.org/abs/2509.11656", "venue": "Conference on Empirical Methods in Natural Language Processing: System Demonstrations (EMNLP Demos-25)", "categories": [ "Runtime", "Evaluation", "Open Problems" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv250921848", "title": "Graph of Agents: Principled Long Context Modeling by Emergent Multi-Agent Collaboration", "authors": "Taejong Joo and Shu Ishida and Ivan Sosnovik and Bryan Lim and Sahand Rezaei-Shoshtari and Adam Gaier and Robert Giaquinto", "year": 2025, "url": "https://arxiv.org/abs/2509.21848", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv251007475", "title": "MAPRO: Recasting Multi-Agent Prompt Optimization as Maximum a Posteriori Inference", "authors": "Zheyuan Zhang and Lin Ge and Hongjiang Li and Weicheng Zhu and Chuxu Zhang and Yanfang Ye", "year": 2025, "url": "https://arxiv.org/abs/2510.07475", "venue": "Conference of the European Chapter of the Association for Computational Linguistics (EACL-26)", "categories": [ "Optimization", "Open Problems" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_3_optimization.tex", "sections/7_outlook.tex" ], "method": "MAPRO", "evidence": { "method": "MAPRO", "family": "Optimization", "subfamily": "Prompt", "update": "design", "effect": "task quality", "boundary": "search cost excluded", "quantified": [ "Paired outcome (Q)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv251016635", "title": "MA-SAPO: Multi-Agent Reasoning for Score-Aware Prompt Optimization", "authors": "Wonduk Seo and Juhyeon Lee and Junseo Koh and Wonseok Choi and Hyunjin An and Jian Park and Seunghyun lee and Haihua Chen and Yi Bu", "year": 2025, "url": "https://arxiv.org/abs/2510.16635", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv251026585", "title": "Stop Wasting Your Tokens: Towards Efficient Runtime Multi-Agent Systems", "authors": "Fulin Lin and Shaowen Chen and Ruishan Fang and Hongwei Wang and Tao Lin", "year": 2025, "url": "https://arxiv.org/abs/2510.26585", "venue": "International Conference on Learning Representations (ICLR-26)", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv251102200", "title": "Optimal-Agent-Selection: State-Aware Routing Framework for Efficient Multi-Agent Collaboration", "authors": "Jingbo Wang and Sendong Zhao and Haochun Wang and Yuzheng Fan and Ting Liu", "year": 2025, "url": "https://arxiv.org/abs/2511.02200", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv251208296", "title": "Towards a Science of Scaling Agent Systems", "authors": "Yubin Kim and Ken Gu and Chanwoo Park and Chunjong Park and Samuel Schmidgall and A. Ali Heydari and Yao Yan and Zhihan Zhang and Yuchen Zhuang and Yun Liu and Mark Malhotra and Paul Pu Liang and Hae Won Park and Yuzhe Yang and Xuhai Xu and Yilun Du and Shwetak Patel and Tim Althoff and Daniel McDuff and Xin Liu", "year": 2025, "url": "https://arxiv.org/abs/2512.08296", "venue": "arXiv preprint", "categories": [ "Collaboration Boundary", "Evaluation", "Synthesis", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/3_boundaries.tex", "sections/5_evaluation.tex", "sections/6_synthesis.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv251220845", "title": "MAR:Multi-Agent Reflexion Improves Reasoning Abilities in LLMs", "authors": "Onat Ozer and Yuchen Wang and Grace Wu and Daniel Dosti and Honghao Zhang and Vivi De La Rue", "year": 2025, "url": "https://arxiv.org/abs/2512.20845", "venue": "arXiv preprint", "categories": [ "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260111913", "title": "LSTM-MAS: A Long Short-Term Memory Inspired Multi-Agent System for Long-Context Understanding", "authors": "Yichen Jiang and Jiakang Yuan and Chongjun Tu and Peng Ye and Tao Chen", "year": 2026, "url": "https://arxiv.org/abs/2601.11913", "venue": "arXiv preprint", "categories": [ "Runtime" ], "subfamilies": [ "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260112996", "title": "OFA-MAS: One-for-All Multi-Agent System Topology Design based on Mixture-of-Experts Graph Generative Models", "authors": "Shiyuan Li and Yixin Liu and Yu Zheng and Mei Li and Quoc Viet Hung Nguyen and Shirui Pan", "year": 2026, "url": "https://arxiv.org/abs/2601.12996", "venue": "ACM Web Conference (WWW-26)", "categories": [ "Topology", "Open Problems" ], "subfamilies": [ "Topology / Topology Adaptation" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260119793", "title": "CASTER: Breaking the Cost-Performance Barrier in Multi-Agent Orchestration via Context-Aware Strategy for Task Efficient Routing", "authors": "Shanyv Liu and Xuyang Yuan and Tao Chen and Zijun Zhan and Zhu Han and Danyang Zheng and Weishan Zhang and Shaohua Cao", "year": 2026, "url": "https://arxiv.org/abs/2601.19793", "venue": "arXiv preprint", "categories": [ "Runtime", "Synthesis" ], "subfamilies": [ "Runtime / Routing" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/6_synthesis.tex" ], "method": "CASTER", "evidence": { "method": "CASTER", "family": "Runtime", "subfamily": "Routing", "update": "step", "effect": "dollars", "boundary": "quality tolerance varies", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260121972", "title": "Learning Decentralized LLM Collaboration with Multi-Agent Actor Critic", "authors": "Shuo Liu and Tianle Chen and Ryan Amiri and Christopher Amato", "year": 2026, "url": "https://arxiv.org/abs/2601.21972", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Policy Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260208847", "title": "Dr. MAS: Stable Reinforcement Learning for Multi-Agent LLM Systems", "authors": "Lang Feng and Longtao Zheng and Shuo He and Fuxiang Zhang and Bo An", "year": 2026, "url": "https://arxiv.org/abs/2602.08847", "venue": "arXiv preprint", "categories": [ "Optimization" ], "subfamilies": [ "Optimization / Policy Learning" ], "sourceFiles": [ "sections/4_3_optimization.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260220229", "title": "HieraMAS: Optimizing Intra-Node LLM Mixtures and Inter-Node Topology for Multi-Agent Systems", "authors": "Tianjun Yao and Zhaoyi Li and Zhiqiang Shen", "year": 2026, "url": "https://arxiv.org/abs/2602.20229", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260302630", "title": "MASPOB: Bandit-Based Prompt Optimization for Multi-Agent Systems with Graph Neural Networks", "authors": "Zhi Hong and Qian Zhang and Jiahang Sun and Zhiwei Shang and Mingze Kong and Xiangyi Wang and Yao Shu and Zhongxiang Dai", "year": 2026, "url": "https://arxiv.org/abs/2603.02630", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Optimization", "Open Problems" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_3_optimization.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260315183", "title": "Token Coherence: Adapting MESI Cache Protocols to Minimize Synchronization Overhead in Multi-Agent LLM Systems", "authors": "Vladyslav Parakhin", "year": 2026, "url": "https://arxiv.org/abs/2603.15183", "venue": "arXiv preprint", "categories": [ "Runtime", "Open Problems" ], "subfamilies": [ "Runtime / State" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260319677", "title": "GoAgent: Group-of-Agents Communication Topology Generation for LLM-based Multi-Agent Systems", "authors": "Hongjiang Chen and Xin Zheng and Yixin Liu and Pengfei Jiao and Shiyuan Li and Huan Liu and Zhidong Zhao and Ziqi Xu and Ibrahim Khalil and Shirui Pan", "year": 2026, "url": "https://arxiv.org/abs/2603.19677", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Construction" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260324579", "title": "MARCH: Multi-Agent Reinforced Self-Check for LLM Hallucination", "authors": "Zhuo Li and Yupeng Zhang and Pengyu Cheng and Jiajun Song and Mengyu Zhou and Hao Li and Shujie Hu and Yu Qin and Erchao Zhao and Xiaoxi Jiang and Guanjun Jiang", "year": 2026, "url": "https://arxiv.org/abs/2603.24579", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-26)", "categories": [ "Runtime", "Optimization", "Open Problems" ], "subfamilies": [ "Optimization / Policy Learning", "Runtime / Communication" ], "sourceFiles": [ "sections/4_2_runtime.tex", "sections/4_3_optimization.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260402460", "title": "Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets", "authors": "Dat Tran and Douwe Kiela", "year": 2026, "url": "https://arxiv.org/abs/2604.02460", "venue": "arXiv preprint", "categories": [ "Collaboration Boundary", "Evaluation", "Open Problems" ], "subfamilies": [], "sourceFiles": [ "sections/3_boundaries.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260403143", "title": "TokenDance: Scaling Multi-Agent LLM Serving via Collective KV Cache Sharing", "authors": "Zhuohang Bian and Feiyang Wu and Chengrui Zhang and Hangcheng Dong and Yun Liang and Youwei Zhuo", "year": 2026, "url": "https://arxiv.org/abs/2604.03143", "venue": "arXiv preprint", "categories": [ "Runtime", "Synthesis", "Open Problems" ], "subfamilies": [ "Runtime / Routing", "Runtime / State" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/4_2_runtime.tex", "sections/6_synthesis.tex", "sections/7_outlook.tex" ], "method": "TokenDance", "evidence": { "method": "TokenDance", "family": "Runtime", "subfamily": "State", "update": "serving", "effect": "KV memory, concurrency", "boundary": "workload and hardware specific", "quantified": [ "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260417148", "title": "Graph-of-Agents: A Graph-based Framework for Multi-Agent LLM Collaboration", "authors": "Sukwon Yun and Jie Peng and Pingzhi Li and Wendong Fan and Jie Chen and James Zou and Guohao Li and Tianlong Chen", "year": 2026, "url": "https://arxiv.org/abs/2604.17148", "venue": "International Conference on Learning Representations (ICLR-26)", "categories": [ "Evaluation", "Synthesis" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex", "sections/6_synthesis.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260506623", "title": "MASPO: Joint Prompt Optimization for LLM-based Multi-Agent Systems", "authors": "Zhexuan Wang and Xuebo Liu and Li Wang and Zifei Shan and Yutong Wang and Zhenxi Song and Min Zhang", "year": 2026, "url": "https://arxiv.org/abs/2605.06623", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Optimization", "Open Problems" ], "subfamilies": [ "Optimization / Prompt Optimization" ], "sourceFiles": [ "sections/4_3_optimization.tex", "sections/7_outlook.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260509539", "title": "TacoMAS: Test-Time Co-Evolution of Topology and Capability in LLM-based Multi-Agent Systems", "authors": "Chen Xu and Yicheng Hu and Ruizi Wang and Xinyu Lin and Wenjie Wang and Dongrui Liu and Fuli Feng", "year": 2026, "url": "https://arxiv.org/abs/2605.09539", "venue": "arXiv preprint", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Adaptation" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "TacoMAS", "evidence": { "method": "TacoMAS", "family": "Topology", "subfamily": "Adaptation", "update": "runtime", "effect": "quality and calls", "boundary": "additional test-time calls", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260517359", "title": "Learning Transferable Topology Priors for Multi-Agent LLM Collaboration Across Domains", "authors": "Taolin Zhang and Zijie Zhou and Jiuheng Wan and Tingyuan Hu and Chengyu Wang and Xiaofeng He and Richang Hong", "year": 2026, "url": "https://arxiv.org/abs/2605.17359", "venue": "arXiv preprint", "categories": [ "Topology", "Evaluation", "Open Problems" ], "subfamilies": [ "Topology / Topology Adaptation" ], "sourceFiles": [ "sections/4_1_topology.tex", "sections/5_evaluation.tex", "sections/7_outlook.tex" ], "method": "TopoPrior", "evidence": { "method": "TopoPrior", "family": "Topology", "subfamily": "Adaptation", "update": "prior + query", "effect": "online tokens", "boundary": "reported break-even", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Offline cost included (O)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" }, { "key": "arxiv260601828", "title": "Dynamic Trust-Aware Sparse Communication Topology for LLM-Based Multi-Agent Consensus", "authors": "Wanshuang Gou and Zihan Liu", "year": 2026, "url": "https://arxiv.org/abs/2606.01828", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "austin2021mbpp", "title": "Program Synthesis with Large Language Models", "authors": "Jacob Austin and Augustus Odena and Maxwell Nye and Maarten Bosma and Henryk Michalewski and David Dohan and Ellen Jiang and Carrie Cai and Michael Terry and Quoc Le and Charles Sutton", "year": 2021, "url": "https://arxiv.org/abs/2108.07732", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "bai2023longbench", "title": "LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding", "authors": "Yushi Bai and Xin Lv and Jiajie Zhang and Hongchang Lyu and Jiankai Tang and Zhidian Huang and Zhengxiao Du and Xiao Liu and Aohan Zeng and Lei Hou and Yuxiao Dong and Jie Tang and Juanzi Li", "year": 2024, "url": "https://arxiv.org/abs/2308.14508", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "carroll2019overcooked", "title": "On the Utility of Learning about Humans for Human-AI Coordination", "authors": "Micah Carroll and Rohin Shah and Mark K. Ho and Thomas L. Griffiths and Sanjit A. Seshia and Pieter Abbeel and Anca Dragan", "year": 2019, "url": "https://arxiv.org/abs/1910.05789", "venue": "Conference on Neural Information Processing Systems (NeurIPS-19)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "chen2021humaneval", "title": "Evaluating Large Language Models Trained on Code", "authors": "Mark Chen and Jerry Tworek and Heewoo Jun and Qiming Yuan and Henrique Ponde de Oliveira Pinto and Jared Kaplan and Harri Edwards and Yuri Burda and Nicholas Joseph and Greg Brockman and Alex Ray and Raul Puri and Gretchen Krueger and Michael Petrov and Heidy Khlaaf and Girish Sastry and Pamela Mishkin and Brooke Chan and Scott Gray and Nick Ryder and Mikhail Pavlov and Alethea Power and Lukasz Kaiser and Mohammad Bavarian and Clemens Winter and Philippe Tillet and Felipe Petroski Such and Dave Cummings and Matthias Plappert and Fotios Chantzis and Elizabeth Barnes and Ariel Herbert-Voss and William Hebgen Guss and Alex Nichol and Alex Paino and Nikolas Tezak and Jie Tang and Igor Babuschkin and Suchir Balaji and Shantanu Jain and William Saunders and Christopher Hesse and Andrew N. Carr and Jan Leike and Josh Achiam and Vedant Misra and Evan Morikawa and Alec Radford and Matthew Knight and Miles Brundage and Mira Murati and Katie Mayer and Peter Welinder and Bob McGrew and Dario Amodei and Sam McCandlish and Ilya Sutskever and Wojciech Zaremba", "year": 2021, "url": "https://arxiv.org/abs/2107.03374", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "clark2018arc", "title": "Think You Have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge", "authors": "Peter Clark and Isaac Cowhey and Oren Etzioni and Tushar Khot and Ashish Sabharwal and Carissa Schoenick and Oyvind Tafjord", "year": 2018, "url": "https://arxiv.org/abs/1803.05457", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "cobbe2021gsm8k", "title": "Training Verifiers to Solve Math Word Problems", "authors": "Karl Cobbe and Vineet Kosaraju and Mohammad Bavarian and Mark Chen and Heewoo Jun and Lukasz Kaiser and Matthias Plappert and Jerry Tworek and Jacob Hilton and Reiichiro Nakano and Christopher Hesse and John Schulman", "year": 2021, "url": "https://arxiv.org/abs/2110.14168", "venue": "arXiv preprint", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "geva2021strategyqa", "title": "Did Aristotle Use a Laptop? A Question Answering Benchmark with Implicit Reasoning Strategies", "authors": "Mor Geva and Daniel Khashabi and Elad Segal and Tushar Khot and Dan Roth and Jonathan Berant", "year": 2021, "url": "https://arxiv.org/abs/2101.02235", "venue": "Transactions of the Association for Computational Linguistics (TACL), 2021", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "guo2024mas_survey", "title": "Large Language Model based Multi-Agents: A Survey of Progress and Challenges", "authors": "Taicheng Guo and Xiuying Chen and Yaqi Wang and Ruidi Chang and Shichao Pei and N. Chawla and Olaf Wiest and Xiangliang Zhang", "year": 2024, "url": "https://arxiv.org/abs/2402.01680", "venue": "International Joint Conference on Artificial Intelligence (IJCAI-24)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "han2024mas_challenges", "title": "LLM Multi-Agent Systems: Challenges and Open Problems", "authors": "Shanshan Han and Qifan Zhang and Yuhang Yao and Weizhao Jin and Zhaozhuo Xu and Chaoyang He", "year": 2024, "url": "https://arxiv.org/abs/2402.03578", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "he2024olympiadbench", "title": "OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems", "authors": "Chaoqun He and Renjie Luo and Yuzhuo Bai and Shengding Hu and Zhen Leng Thai and Junhao Shen and Jinyi Hu and Xu Han and Yujie Huang and Yuxiang Zhang and Jie Liu and Lei Qi and Zhiyuan Liu and Maosong Sun", "year": 2024, "url": "https://arxiv.org/abs/2402.14008", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "hendrycks2021apps", "title": "Measuring Coding Challenge Competence With APPS", "authors": "Dan Hendrycks and Steven Basart and Saurav Kadavath and Mantas Mazeika and Akul Arora and Ethan Guo and Collin Burns and Samir Puranik and Horace He and Dawn Song and Jacob Steinhardt", "year": 2021, "url": "https://arxiv.org/abs/2105.09938", "venue": "Conference on Neural Information Processing Systems (NeurIPS-21), Datasets and Benchmarks Track", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "hendrycks2021math", "title": "Measuring Mathematical Problem Solving With the MATH Dataset", "authors": "Dan Hendrycks and Collin Burns and Saurav Kadavath and Akul Arora and Steven Basart and Eric Tang and Dawn Song and Jacob Steinhardt", "year": 2021, "url": "https://arxiv.org/abs/2103.03874", "venue": "Conference on Neural Information Processing Systems (NeurIPS-21), Datasets and Benchmarks Track", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "hendrycks2021mmlu", "title": "Measuring Massive Multitask Language Understanding", "authors": "Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt", "year": 2021, "url": "https://arxiv.org/abs/2009.03300", "venue": "International Conference on Learning Representations (ICLR-21)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "ho2020_2wiki", "title": "Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Steps", "authors": "Xanh Ho and Anh-Khoa Duong Nguyen and Saku Sugawara and Akiko Aizawa", "year": 2020, "url": "https://arxiv.org/abs/2011.01060", "venue": "International Conference on Computational Linguistics (COLING-20)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "jain2024livecodebench", "title": "LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code", "authors": "Naman Jain and King Han and Alex Gu and Wen-Ding Li and Fanjia Yan and Tianjun Zhang and Sida Wang and Armando Solar-Lezama and Koushik Sen and Ion Stoica", "year": 2025, "url": "https://arxiv.org/abs/2403.07974", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "jimenez2024swebench", "title": "SWE-bench: Can Language Models Resolve Real-World GitHub Issues?", "authors": "Carlos E. Jimenez and John Yang and Alexander Wettig and Shunyu Yao and Kexin Pei and Ofir Press and Karthik Narasimhan", "year": 2024, "url": "https://arxiv.org/abs/2310.06770", "venue": "International Conference on Learning Representations (ICLR-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "joshi2017triviaqa", "title": "TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension", "authors": "Mandar Joshi and Eunsol Choi and Daniel S. Weld and Luke Zettlemoyer", "year": 2017, "url": "https://arxiv.org/abs/1705.03551", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-17)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "kong2025token", "title": "Token Reduction Should Go Beyond Efficiency in Generative Models – From Vision, Language to Multimodality", "authors": "Zhenglun Kong and Yize Li and Fanhu Zeng and Lei Xin and Shvat Messica and Xue Lin and Pu Zhao and Manolis Kellis and Hao Tang and Marinka Zitnik", "year": 2025, "url": "https://arxiv.org/abs/2505.18227", "venue": "arXiv", "categories": [ "Background" ], "subfamilies": [ "Runtime / Communication", "Runtime / State" ], "metadataSource": "https://arxiv.org/abs/2505.18227v4", "collectionSource": "Project-page addition", "sourceFiles": [], "method": "", "evidence": null }, { "key": "kwiatkowski2019nq", "title": "Natural Questions: A Benchmark for Question Answering Research", "authors": "Tom Kwiatkowski and Jennimaria Palomaki and Olivia Redfield and Michael Collins and Ankur Parikh and Chris Alberti and Danielle Epstein and Illia Polosukhin and Jacob Devlin and Kenton Lee and Kristina Toutanova and Llion Jones and Matthew Kelcey and Ming-Wei Chang and Andrew M. Dai and Jakob Uszkoreit and Quoc Le and Slav Petrov", "year": 2019, "url": "https://aclanthology.org/Q19-1026/", "venue": "Transactions of the Association for Computational Linguistics (TACL), 2019", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "li2022codecontests", "title": "Competition-Level Code Generation with AlphaCode", "authors": "Yujia Li and David Choi and Junyoung Chung and Nate Kushman and Julian Schrittwieser and Rémi Leblond and Tom Eccles and James Keeling and Felix Gimeno and Agustin Dal Lago and Thomas Hubert and Peter Choy and Cyprien de Masson d'Autume and Igor Babuschkin and Xinyun Chen and Po-Sen Huang and Johannes Welbl and Sven Gowal and Alexey Cherepanov and James Molloy and Daniel J. Mankowitz and Esme Sutherland Robson and Pushmeet Kohli and Nando de Freitas and Koray Kavukcuoglu and Oriol Vinyals", "year": 2022, "url": "https://arxiv.org/abs/2203.07814", "venue": "Science, 2022", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "li2026sponge", "title": "Sponge tool attack: Stealthy denial-of-efficiency against tool-augmented agentic reasoning", "authors": "Qi Li and Xinchao Wang", "year": 2026, "url": "https://arxiv.org/abs/2601.17566", "venue": "International Conference on Machine Learning (ICML-26)", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "ling2017aqua", "title": "Program Induction by Rationale Generation: Learning to Solve and Explain Algebraic Word Problems", "authors": "Wang Ling and Dani Yogatama and Chris Dyer and Phil Blunsom", "year": 2017, "url": "https://arxiv.org/abs/1705.04146", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-17)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "liu2021logiqa", "title": "LogiQA: A Challenge Dataset for Machine Reading Comprehension with Logical Reasoning", "authors": "Jian Liu and Leyang Cui and Hanmeng Liu and Dandan Huang and Yile Wang and Yue Zhang", "year": 2020, "url": "https://arxiv.org/abs/2007.08124", "venue": "International Joint Conference on Artificial Intelligence (IJCAI-20)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "maharana2024locomo", "title": "Evaluating Very Long-Term Conversational Memory of LLM Agents", "authors": "Adyasha Maharana and Dong-Ho Lee and Sergey Tulyakov and Mohit Bansal and Francesco Barbieri and Yuwei Fang", "year": 2024, "url": "https://arxiv.org/abs/2402.17753", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "patel2021svamp", "title": "Are NLP Models Really Able to Solve Simple Math Word Problems?", "authors": "Arkil Patel and Satwik Bhattamishra and Navin Goyal", "year": 2021, "url": "https://arxiv.org/abs/2103.07191", "venue": "Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT-21)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "rein2023gpqa", "title": "GPQA: A Graduate-Level Google-Proof Q&A Benchmark", "authors": "David Rein and Betty Li Hou and Asa Cooper Stickland and Jackson Petty and Richard Yuanzhe Pang and Julien Dirani and Julian Michael and Samuel R. Bowman", "year": 2024, "url": "https://arxiv.org/abs/2311.12022", "venue": "Conference on Language Modeling (COLM-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "roy2015multiarith", "title": "Solving General Arithmetic Word Problems", "authors": "Subhro Roy and Dan Roth", "year": 2015, "url": "https://aclanthology.org/D15-1202/", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-15)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "shi2025aime", "title": "Aime: Towards Fully-Autonomous Multi-Agent Framework", "authors": "Yexuan Shi and Mingyu Wang and Yunxiang Cao and Hongjie Lai and Junjian Lan and Xin Han and Yu Wang and Jie Geng and Zhenan Li and Zihao Xia and Xiang Chen and Chen Li and Jian Xu and Wenbo Duan and Yuanshuo Zhu", "year": 2025, "url": "https://arxiv.org/abs/2507.11988", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "shridhar2020alfworld", "title": "ALFWorld: Aligning Text and Embodied Environments for Interactive Learning", "authors": "Mohit Shridhar and Xingdi Yuan and Marc-Alexandre Côté and Yonatan Bisk and Adam Trischler and Matthew Hausknecht", "year": 2021, "url": "https://arxiv.org/abs/2010.03768", "venue": "International Conference on Learning Representations (ICLR-21)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "sprague2024musr", "title": "MuSR: Testing the Limits of Chain-of-Thought with Multistep Soft Reasoning", "authors": "Zayne Sprague and Xi Ye and Kaj Bostrom and Swarat Chaudhuri and Greg Durrett", "year": 2024, "url": "https://arxiv.org/abs/2310.16049", "venue": "International Conference on Learning Representations (ICLR-24), Spotlight", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "suzgun2022bbh", "title": "Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them", "authors": "Mirac Suzgun and Nathan Scales and Nathanael Sch\"arli and Sebastian Gehrmann and Yi Tay and Hyung Won Chung and Aakanksha Chowdhery and Quoc V. Le and Ed H. Chi and Denny Zhou and Jason Wei", "year": 2023, "url": "https://arxiv.org/abs/2210.09261", "venue": "Annual Meeting of the Association for Computational Linguistics (Findings of ACL-23)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "talmor2019csqa", "title": "CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge", "authors": "Alon Talmor and Jonathan Herzig and Nicholas Lourie and Jonathan Berant", "year": 2019, "url": "https://arxiv.org/abs/1811.00937", "venue": "Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT-19)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "thorne2018fever", "title": "FEVER: A Large-Scale Dataset for Fact Extraction and VERification", "authors": "James Thorne and Andreas Vlachos and Christos Christodoulopoulos and Arpit Mittal", "year": 2018, "url": "https://arxiv.org/abs/1803.05355", "venue": "Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT-18)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "tran2025masaboration", "title": "Multi-Agent Collaboration Mechanisms: A Survey of LLMs", "authors": "Khanh-Tung Tran and Dung Dao and Minh-Duong Nguyen and Quoc-Viet Pham and Barry O'Sullivan and Hoang D. Nguyen", "year": 2025, "url": "https://arxiv.org/abs/2501.06322", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "trivedi2022musique", "title": "MuSiQue: Multihop Questions via Single-hop Question Composition", "authors": "Harsh Trivedi and Niranjan Balasubramanian and Tushar Khot and Ashish Sabharwal", "year": 2022, "url": "https://arxiv.org/abs/2108.00573", "venue": "Transactions of the Association for Computational Linguistics (TACL), 2022", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "trivedi2024appworld", "title": "AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents", "authors": "Harsh Trivedi and Tushar Khot and Mareike Hartmann and Ruskin Manku and Vinty Dong and Edward Li and Shashank Gupta and Ashish Sabharwal and Niranjan Balasubramanian", "year": 2024, "url": "https://arxiv.org/abs/2407.18901", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "wang2022scienceworld", "title": "ScienceWorld: Is Your Agent Smarter than a 5th Grader?", "authors": "Ruoyao Wang and Peter Jansen and Marc-Alexandre Côté and Prithviraj Ammanabrolu", "year": 2022, "url": "https://arxiv.org/abs/2203.07540", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-22)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "wang2024mmlupro", "title": "MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark", "authors": "Yubo Wang and Xueguang Ma and Ge Zhang and Yuansheng Ni and Abhranil Chandra and Shiguang Guo and Weiming Ren and Aaran Arulraj and Xuan He and Ziyan Jiang and Tianle Li and Max Ku and Kai Wang and Alex Zhuang and Rongqi Fan and Xiang Yue and Wenhu Chen", "year": 2024, "url": "https://arxiv.org/abs/2406.01574", "venue": "Conference on Neural Information Processing Systems (NeurIPS-24), Datasets and Benchmarks Track", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "xie2024travelplanner", "title": "TravelPlanner: A Benchmark for Real-World Planning with Language Agents", "authors": "Jian Xie and Kai Zhang and Jiangjie Chen and Tinghui Zhu and Renze Lou and Yuandong Tian and Yanghua Xiao and Yu Su", "year": 2024, "url": "https://arxiv.org/abs/2402.01622", "venue": "International Conference on Machine Learning (ICML-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "yang2018hotpotqa", "title": "HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering", "authors": "Zhilin Yang and Peng Qi and Saizheng Zhang and Yoshua Bengio and William W. Cohen and Ruslan Salakhutdinov and Christopher D. Manning", "year": 2018, "url": "https://arxiv.org/abs/1809.09600", "venue": "Conference on Empirical Methods in Natural Language Processing (EMNLP-18)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "yao2022webshop", "title": "WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents", "authors": "Shunyu Yao and Howard Chen and John Yang and Karthik Narasimhan", "year": 2022, "url": "https://arxiv.org/abs/2207.01206", "venue": "Conference on Neural Information Processing Systems (NeurIPS-22)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "zhang2024infinitebench", "title": "∞Bench: Extending Long Context Evaluation Beyond 100K Tokens", "authors": "Xinrong Zhang and Yingfa Chen and Shengding Hu and Zihang Xu and Junhao Chen and Moo Khai Hao and Xu Han and Zhen Leng Thai and Shuo Wang and Zhiyuan Liu and Maosong Sun", "year": 2024, "url": "https://arxiv.org/abs/2402.13718", "venue": "Annual Meeting of the Association for Computational Linguistics (ACL-24)", "categories": [ "Evaluation" ], "subfamilies": [], "sourceFiles": [ "sections/5_evaluation.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "zhang2025agentorchestra", "title": "AgentOrchestra: Orchestrating Multi-Agent Intelligence with the Tool-Environment-Agent(TEA) Protocol", "authors": "Wentao Zhang and Liang Zeng and Yuzhen Xiao and Yongcong Li and Ce Cui and Yilei Zhao and Rui Hu and Yang Liu and Yahui Zhou and Bo An", "year": 2025, "url": "https://arxiv.org/abs/2506.12508", "venue": "arXiv preprint", "categories": [ "Introduction" ], "subfamilies": [], "sourceFiles": [ "sections/1_intro.tex" ], "method": "", "evidence": null, "metadataSource": "Manuscript bibliography" }, { "key": "zhang2025agentprune", "title": "Cut the Crap: An Economical Communication Pipeline for LLM-based Multi-Agent Systems", "authors": "Guibin Zhang and Yanwei Yue and Zhixun Li and Sukwon Yun and Guancheng Wan and Kun Wang and Dawei Cheng and Jeffrey Xu Yu and Tianlong Chen", "year": 2024, "url": "https://arxiv.org/abs/2410.02506", "venue": "International Conference on Learning Representations (ICLR-25)", "categories": [ "Topology" ], "subfamilies": [ "Topology / Topology Pruning" ], "sourceFiles": [ "sections/4_1_topology.tex" ], "method": "AgentPrune", "evidence": { "method": "AgentPrune", "family": "Topology", "subfamily": "Pruning", "update": "task", "effect": "prompt tokens, dollars", "boundary": "offline graph search", "quantified": [ "Paired outcome (Q)", "Measured deployment use (R)", "Cross-setting evaluation (X)" ] }, "metadataSource": "Manuscript bibliography" } ] }