Tag
Introduces RACL, a reasoning-agent control layer that improves metaheuristic optimization by learning to control internal search behavior from operational memory, showing cost improvements in vehicle routing tests.