Dynamic Parameters Ant Colony Algorithm with Particle Swarm Characteristic

Abstract

To improve the convergence time of ant colony algorithm, avoid falling in local best and enhance the quality of solution, a novel dynamic parameters ant colony algorithm with particle swarm characteristics is proposed. Learning the multi-information instruction characteristic of Particle Swarm Optimization Algorithm, the global pheromone update rule with particle swarm characteristic is introduced to improve the directive function of pheromone and the speed of convergence. At the same time, solution multiplicity is guaranteed as far as possible. Using the function of current condition to update particle speed and position, parameters of Ant Colony Algorithm is used to reflect the current condition. Hyperbola Tangent function is imported to dynamic adjust parameters so that the relation between local search and global search could be balanced. Comparing with basic Ant Colony Algorithm, the simulation result on TSP shows that new algorithm has higher convergence speed and better solution.