Ant Colony Optimization with Policy Gradients and Replay

Author(s): Jardee, W., & Sheppard, J.

Abstract

Ant Colony Optimization (ACO) has served as a widely-utilized metaheuristic algorithm for decades for solving combinatorial optimization problems. Since its initial construction, ACO has seen a wide variety of modifications and connections to Reinforcement Learning (RL). Substantial parallels can be seen as early as 1995 with Ant-Q’s relationship with Q-learning, through 2022 with ADACO’s connection with Policy Gradient. In this work, we describe ACO, more specifically the Stochastic Gradient Descent ACO algorithm (ACOSGD), explicitly as an off-policy Policy Gradient (PG) method. We also incorporate experience replay into several ACO algorithm variants, including AS, MaxMin-ACO, ACOSGD, ADACO, and our two policy gradient-based versions: PGACO and PPOACO, drawing the connection to elitist ACO strategies. We show that our implementation of PG in ACO with experience replay and a baselined reward update strategy applied to eight TSP problems of varying sizes performs competitively with both fundamental ACO and SGD-based ACO versions. We also show that the replay buffer seems to unilaterally improve the performance of ACO algorithms through an ablation study.

Citation

Jardee, W., & Sheppard, J. (2025). Ant colony optimization with policy gradients and replay. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO ’25) (pp. 240–248). Association for Computing Machinery. https://doi.org/10.1145/3712256.3726452