Improved Robustness through Population Variance in Ant Colony Optimization

Ant Colony Optimization algorithms are population-based Stochastic Local Search algorithms that mimic the behavior of ants, simulating pheromone trails to search for solutions to combinatorial optimization problems. This paper introduces Population Variance, a novel approach to ACO algorithms that allows parameters to vary across the population over time, leading to solution construction differences that are not strictly stochastic. The increased exploration appears to help the search escape from local optima, significantly improving the robustness of the algorithm with respect to suboptimal parameter settings.