Multi-Agent Planning

Follow this topic Ask about this topic

Multi-Agent Planning studies how multiple robots, vehicles, software agents, or people can coordinate actions to meet shared or competing goals. Methods include search and optimization for task assignment and pathfinding, decentralized planning with communication, reinforcement learning and world models, game-theoretic planning, and language-model agents organized into roles and workflows. Work is evaluated in settings such as warehouse routing, robot navigation, autonomous driving, games, and long-horizon computer-use tasks, with measures of success, safety, efficiency, and coordination. A central challenge is balancing scalability and responsiveness with reliable coordination under partial information, changing conditions, and limited communication, while preserving formal guarantees where needed.

Latest papers

121 papers20 papers in the last four weeks, up 233% on the four weeks before. 0.1% of all new papers.

GraphChase: A Platform and Benchmark for Urban Network Security Games
Jan 29, 2025Shuxin Zhuang, Shuxin Li, Tianji Yang +4Benchmark Design AI Agent Benchmarks
City University of Hong Kong CAIR, Hong Kong Institute of Science & Innovation, CAS Hong Kong, Hong Kong · Nanyang Technological University Singapore, Singapore · Georgia Institute of Technology Atlanta, Georgia, United States +2

Jan 29, 2025J/K move · Enter open · S save

GraphChase: A Platform and Benchmark for Urban Network Security Games

Shuxin Zhuang, Shuxin Li, Tianji Yang, Muheng Li+3

City University of Hong Kong CAIR, Hong Kong Institute of Science & Innovation, CAS Hong Kong, Hong Kong · Nanyang Technological University Singapore, Singapore · Georgia Institute of Technology Atlanta, Georgia, United States+2

Benchmark Design AI Agent Benchmarks Multi-Agent Planning

Open paper AskPDF arXiv

After the achievement of solving two-player zero-sum games, more AI researchers focus on solving multiplayer games. Urban Network Security Games (UNSGs) represent a class of such games, modeling real-world scenarios where law enforcement must strategically allocate limited resources to intercept criminals escaping within urban networks, and have gained considerable research attention. However, progress in this field has been limited by the absence of a standardized experimental platform and realistic benchmarks with heterogeneous travel costs. To address this limitation, we introduce GraphChase, an open-source platform designed to support the development and evaluation of algorithms for UNSGs. GraphChase offers a unified environment for modeling diverse UNSG variants on unweighted and weighted road networks across urban topologies. It also incorporates learning-based algorithms as baseline references for researchers. Furthermore, our experiments with GraphChase reveal that existing approaches to UNSGs still face challenges in terms of robustness and scalability, and suffer performance degradation when deployed under weighted edge costs, highlighting a sim-to-real generalization gap. GraphChase thus provides a realistic testbed for developing and validating UNSGs solvers under realistic travel-time heterogeneity.