AI RESEARCH
Taming "Zombie'' Agents: A Markov State-Aware Framework for Resilient Multi-Agent Evolution
arXiv CS.CL
•
ArXi:2605.17348v1 Announce Type: new Recent advancements in LLM-based multi-agent systems have nstrated remarkable collaborative capabilities across complex tasks. To improve overall efficiency, existing methods often rely on aggressive graph evolution among agents (e.g., node or edge pruning), which risks prematurely discarding valuable agents due to transient issues such as hallucinations or temporary knowledge gaps. However, such hard pruning overlooks the potential for ``zombie'' agents to recover and contribute in subsequent discussion rounds.