AgentRevive is a Markov state-aware framework with policy learning and edge optimization that manages Active, Standby, and Terminated agent states to enable resilient multi-agent evolution and reduce token consumption.
CoRR , volume =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Taming "Zombie'' Agents: A Markov State-Aware Framework for Resilient Multi-Agent Evolution
AgentRevive is a Markov state-aware framework with policy learning and edge optimization that manages Active, Standby, and Terminated agent states to enable resilient multi-agent evolution and reduce token consumption.