Beyond Static Priors: Dynamic Neural Guidance for Large-Scale Ant Colony Optimization

Dat Thanh Tran, Van Khu Vu, Yining Ma

Neural-guided Ant Colony Optimization (ACO) suffers from a fundamental training-inference misalignment: policies are typically trained to generate static priors (e.g., heatmaps), y…