Ant Colony Algorithm For Parameter Optimization And Its Applications

As a new bionic optimized algorithm, ant colony algorithm is a complex intelligent system in essence, which has the advantages of strong robustness, excellent distribution mechanism and convenient integration with other algorithms. It has been successfully applied in many fields. But, there still are many unsolved problems in the theories of ant colony algorithm. Such as: the convergence rate, the pheromone allocation and the parameter selection, among which the parameter selection affects the performance of ant colony algorithm most. Therefore this paper mainly researches the optimization of the parameter selection for ant colony algorithm.In this paper, the performance, principle and mechanism of basic ant colony algorithm are analyzed comprehensively. The effect of the parameter selection on the performance of ant colony algorithm is thoroughly discussed by the simulation experiments of the traveling salesman problem (TSP). Especially the three parameters that have more affect on the algorithm, which are the heuristic factor a, the expectation heuristic factorβ, and the information persistent factorρ. Based on the particle swarm optimization (PSO) algorithm, a parameter optimization scheme is designed to effectively improve the performance of the algorithm by optimizing the combination of the three parameters. Besides, a JAVA program aiming at the Oliver 30 cities problem is written, which generates ideal results. This scheme breaks the traditional limitations which determine the parameters by the experiences and the instincts. Instead, it plays the associated effects of the three parameters, which enables better effects while using ant colony algorithm in the practical optimization problems.In order to verify the feasibility and the practicability of the scheme, it is applied to the Job Shop Scheduling Problems (JSSP). This paper makes an overall summarization on the common methods of JSSP, analyzes the prototype of JSSP, and establishes a new JSSP model based on ant colony algorithm, for the convenience of programming. In the end two typical problems, JSSP6*6 and JSSP10*10, are simulated and the prospective results are obtained.In summary, the parameter optimization scheme in this paper is practicable, and the simulation results are ideal. It can provide some references for the related research works on the ant colony algorithm.