arXiv 论文提出智能体执行计划的修复、重规划与放弃阈值理论
When AI Agents Should Repair, Replan, or Abandon: A Threshold Theory of Plan Commitment
做智能体的朋友值得看这篇:它把“动作挂了该修还是该重排”做成了带阈值的数学决策,还证明重规划成本趋零后稳定性价格才是计划承诺的来源,比工程里拍的 retry 规则有依据得多。
一篇 arXiv 论文(编号 2611.19663)针对智能体在执行中发现计划失效后的决策问题,推导出最优策略:以“有效破坏深度”的阈值决定修复、重规划或放弃目标。论文将修复定义为保留计划结构的部分重置,重规划为全量重置,且修改计划有代价,因为其他参与者已基于该计划做出反应。
核心结论包括:单个失效动作在期望上需要重新验证 1+R(G)/n 个动作,其中 R(G) 统计由依赖路径连接的动作对;完全串行计划切换到重规划所需的原始破坏深度,约比 w 条并行链的计划小 w 倍。论文还给出闭式“稳定性价格”阈值,判断何时保留修复后的较差计划优于采用更好的新计划。当语言模型规划器把重规划成本压到接近零时,稳定性价格成为维持计划承诺的唯一力量;在反复破坏下,修复会累积“补丁债务”,直到债务超过修复的成本优势。
作者用独立实现的数值检验器验证全部命题,160 万次静态模型评估零违规,且模型的区间划分图能定性地复现既有基准中的策略切换结果。论文同时证明了“诊断价值”条件:当破坏可预期地很深且重规划是更优退路时,“总是重规划”才是最优策略。
When AI Agents Should Repair, Replan, or Abandon: A Threshold Theory of Plan Commitment
AI agents often discover mid-execution that their plan is broken, for example when an action fails or a tool is withdrawn. Deployed systems use conventions such as retry loops or "always replan," and a recent benchmark finds no fixed strategy dominates. This paper derives the optimal policy. In the model, a disruption destroys the causal support of part of a plan; repair is a partial reset that preserves the plan's structure, replanning is a full reset, and changing a plan is costly because others have conditioned on it. The optimal policy is a threshold in effective disruption depth: repair shallow damage, replan after deep damage, and abandon the goal when even a fresh plan is not worth its cost. Depth depends on plan structure. In expectation, a single broken action forces 1+R(G)/n actions to be revalidated, where R(G) counts the pairs of actions joined by a dependency path; to first order, a fully serial plan switches to replanning at raw damage about w times smaller than a plan of w parallel chains. A closed-form stability-price threshold determines when a worse repaired plan should be kept over a better fresh one. As language-model planners push replanning costs toward zero, the stability price becomes the only force sustaining plan commitment. Under repeated disruptions, revision frictions make commitment hysteresis optimal: the agent keeps executing a plan it would never choose fresh, and each repair adds patch debt until the debt outweighs repair's cost advantage and the agent rebuilds. A value-of-diagnosis result states when measuring damage is worth its compute and shows that "always replan" is optimal when damage is predictably deep and replanning is the better fallback. A separately implemented numerical checker tests each proposition (zero violations; 1.6 million static-model evaluations), and the model's regime map reproduces that benchmark's regime flips qualitatively.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include agents, arXiv, papers.
Source attribution: curated from arXiv cs.AI (@Gaurav Tewari).