Multi-Agent Systems
Overview
Multi-Agent Systems studies how multiple decision-makers interact when they share goals, compete for resources, or have only partial information about one another. Approaches include multi-agent reinforcement learning, game theory and formal verification, decentralized consensus and task allocation, and—more recently—LLM-agent teams organized through roles, workflows, debate, and shared memory. Researchers evaluate these methods in social dilemmas, cooperative robotics and swarms, reasoning and coding benchmarks, and operational settings, measuring success alongside cost, robustness, and fairness. A central open tension is that coordination can improve outcomes but also adds overhead and can spread shared errors, so its benefits depend on the task, communication structure, and agents involved.
Latest papers
709 papers61 papers in the last four weeks, down 41% on the four weeks before. 0.5% of all new papers.
Permutation Robustness Is Not Enough: Action Collapse in Multi-Agent Transformer Policies
Inherit-MAS: Test-Time Evolution of Multi-Agent Systems through Workflow and Execution Inheritance
Global Coherence: When Every Agent Is Right and the Team Is Still Wrong - A Local-to-Global Semantic Foundation for Multi-Agent Collaboration
It Takes Workflows to Evolve Better Workflows
Pay for the Fault, Not the Flow: Label-Free In-Flow Multi-Agent Workflow Optimization
Worse Together: How Performance Breaks Down in Multi-User Multi-Agent Teams
Multi-agent Auditory Scene Analysis: Improved Localization Speed and Robustness by Multi-beamformed Speech Quality Feedback
ReWAM: Reciprocal World Action Models for Interactive Autonomous Driving
Where Do Multi-Agent Systems Fail? Evidence-Grounded Diagnosis of Collective Mechanisms
CollabFlow: Recursive Self-Improvement of Agent Collaboration
PANDA: A Decentralized Architecture with Flexible Orchestration for Scalable, Fault-Tolerant Multi-Agent Systems
Absorbing State Phase Transitions in Multi-Agent Search
EpiCon: Collective Agent Learning through Co-Evolving Multimodal Memory
LLM-Based Multi-Agent Systems over Wireless Networks: A Joint Agent--Network Design Perspective
Gödel Forest: Balancing Search Depth and Breadth for Data-Centric Recursive Self-Improvement
Memory in the Sky: Low-Altitude Question Answering with Multi-Agent Memory Aggregation
Persistent Partners Raise Prices Among Learning Agents
Collaborative Principle Evolution via Evidence Transfer for Scientific Discovery
Prompted Identity Degrades Cooperation in Multi-Agent LLM Systems
CEO Arena: Evaluating Long-Horizon Multi-Agent Decision-Making in Competitive Markets
ResonAct: Streaming Metrics for Runtime Diagnosis and Self-Healing in Multi-Agent Systems
LLMs for Executable Multi-Agent System Specification Generation
genRTEC', a method that leverages pre-trained Large Language Models (LLMs) to generate executable MAS specifications, in the language of the Run-Time Event Calculus' (RTEC), from natural language descriptions. genRTEC constructs MAS specifications with complex hierarchical and cyclic dependencies based only on short natural language descriptions of the concepts involved. We present an extensive empirical evaluation of genRTEC, spanning various MAS specifications, including both a qualitative and a quantitative assessment. Our results demonstrate that genRTEC constructs executable MAS specifications of high predictive accuracy without compromising reasoning efficiency.