AgentGA: Evolving Code Solutions in Agent-Seed Space
Introduces a new optimization dimension for autonomous code-search systems, showing that evolving starting conditions can significantly improve performance in tabular AutoML.
AgentGA optimizes the initial conditions (agent seeds) for autonomous code-generation agents, achieving 74.52% Exceeds % of Human on the Weco-Kaggle Lite benchmark, outperforming AIDE's 54.15%. Inherited artifacts from parent archives improve later runs.
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 from a reset 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. On the 10 benchmark runs reported here, AgentGA averages 74.52% Exceeds % of Human versus 54.15% for AIDE. Across 1135 parent-child comparisons, descendants given parent archives outperform runs started from scratch, indicating that inherited artifacts improve later autonomous runs. These findings support agent-seed optimization as a practical design point for autonomous code-search systems.