Study of Β,ρ and Q0 Parameters for Shortest Path Estimation Using Ant Colony Optimization

The potential of Ant Colony Optimization (ACO) is massively dependent on its initialization parameters namely β, ρ and Q 0 . In ACO, β is the weight of the heuristic function, ρ is a parameter on which the pheromone update depends and Q 0 is responsible for exploration of the graph. The presented article deals with a methodology to auto tune the ACO parameters for extracting the optimal performance from ACO. An extensive study has been presented considering different dataset to validate the approach. As a result offered by auto tuned ACO, an error of 0.21% from optimum value for estimating shortest path was achieved. The method presented is organized, systematic and promises to extract the best out of ACO.