Extension of Ant System algorithms with exponential pheromone deposition rule for improved performance

摘要

The paper extends the classical Ant System (AS) algorithms by proposing a novel approach of exponential pheromone deposition by artificial ants ensuring a concentration gradient along solution paths. The stability analysis with a deterministic mathematical model based on differential equation yields the proper range of the parameters. A roadmap of connected cities, where the shortest path between a source-destination pair is to be determined, is taken as a problem environment. Exhaustive simulations confirm that the proposed deposition rule, with properly chosen parameter values, outperforms the traditional one with large margin both in terms of solution quality and algorithm convergence.