RACL: Reasoning-Agent Control Layers for Continuous Metaheuristic Learning
This paper introduces RACL, a Reasoning-Agent Control Layer designed to enhance metaheuristic learning. RACL operates by placing a reasoning agent above an existing optimizer, influencing its search behavior through observation, hypothesis formulation, and policy consolidation. Experiments using the vehicle routing problem demonstrate RACL's effectiveness, showing improvements over existing policies and minimal computational overhead. The system utilized Codex as an in-the-loop reasoning agent during its proof-of-concept phase. AI
IMPACT Introduces a novel method for improving optimization algorithms through agent-based reasoning, potentially impacting fields requiring complex problem-solving.