Anchor and Perturb: Lazy Agent Remediation by Exploration Injection

2026-09-23

AI Summary

AnP框架通过向表现不佳的代理注入探索性脉冲，同时保持其他代理的稳定策略，解决了多代理协调失败问题，无需修改网络结构。

Abstract

Anchor and Perturb (AnP) is a lightweight framework that resolves multi-agent coordination failures by decoupling exploratory variance injection from recurrent manifold stability. Existing remediation strategies predominantly alter mixing network architectures or enforce simultaneous exploration across the collective, which inevitably precipitates severe temporal-difference penalties in non-monotonic reward spaces. Specifically, AnP isolates underperforming lazy agents and injects an asymmetric exploratory pulse into targeted coordinates whilst anchoring converged teammates to nominal greedy exploitation. Empirical telemetry benchmarks demonstrate that AnP successfully rescues collapsed joint policies (recovering from a 5% evaluation win rate nadir back to 85%) and facilitates escape from suboptimal coordination plateaus, sustaining peak win rates of 90% without requiring structural network modifications.

Problem

• multi-agent coordination
• lazy agents
• exploration injection
• suboptimal coordination

Innovation

• Exploratory Variance Injection
• Asymmetric Exploratory Pulse
• Lazy Agents
• Temporal-Difference Penalties
• Multi-Agent Coordination