通过优化代码生成的起始条件，让智能体自动进化出更优解。
AgentGA: Evolving Code Solutions in Agent-Seed Space
- 用遗传算法优化任务提示和初始档案，代替直接改代码。
- 在16个竞赛中平均超越人类71.9%，胜率15/16，显著优于基线。
- 继承父代成果的子代胜率51.9%，远超全新方案的8.6%。
我们提出AgentGA框架，通过优化代理种子（任务提示及可选父代档案）来演化自主代码生成过程。外层循环搜索可复用的起始条件，而非直接修改代码。每一代在隔离环境中启动全新自主运行，而选定的父代档案提供可被后代查看与重用的遗产。AgentGA结合种群级遗传算法与长时程智能体；选择采用确定性1:1精英锦标赛，操作符分配通过改进的Hedge控制器在线调整。我们在Weco-Kaggle Lite基准的16个表格式AutoML竞赛中实现该方法。整体上，AgentGA平均超越人类71.90%，高于基线AIDE的51.38%，赢得15/16项竞赛。在内部实验中，基于继承父代档案的后代在1,680场父子对战中获胜51.9%，而全新提案仅8.6%。结果表明，代理种子优化是自主代码搜索系统的一项实用设计选择。
原文摘要 · Abstract (English)
We present AgentGA, a framework that evolves autonomous code-generation runs by optimizing the agent seed: the task prompt plus optional parent archives that initialize a fresh workspace. The outer loop searches over these reusable starting conditions rather than editing code directly. Each generation launches a fresh autonomous run in an isolated workspace, while selected parent archives provide inherited artifacts that descendants can inspect and reuse. AgentGA couples a population-level genetic algorithm with long-horizon agents; selection uses deterministic 1:1 elite tournaments and operator allocation is adapted online with a modified Hedge controller. We instantiate the approach for tabular AutoML on the 16-competition Weco-Kaggle Lite benchmark. Across the full benchmark, AgentGA averages 71.90% Exceeds % of Human versus 51.38% for the AIDE reference, winning 15/16 competitions. Within AgentGA runs, descendants conditioned on inherited parent archives win 51.9% of 1,680 parent-child tournaments versus 8.6% for de novo proposals. These results support agent-seed optimization as a practical design choice for autonomous code-search systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。