Ant Colony Algorithm Research And Its Application

Among heuristic algorithms, Ant Colony Optimization Algorithm(ACO) is being more and more attention, because it has the advantages of strong robustness, excellent distribution mechanism and convenient integration with other algorithms. It has been successfully applied in many fields. But the ACO also has many deficiencies, such as easy prone to local optimum, long running time and parameters'choose is too difficult and so on. For the deficiencies of ACO, this paper carry out research on improved ACO. For the difficult of choosing the parameters of ACO, this paper carry out research on the combinatorial optimization of ACO's parameters. Proposed ant colony algorithm parameters' combination optimization program based on Particle Swarm Optimization Algorithm(PSO).Designed parameters training system.Furthermore, For simulation system of ACO for TSP problem still blank, design and implementation of ACO simulation system based on multi-thread. The work of this paper as follow:First: Elaborate the background and significance of this paper. Summarise the common improved ant colony algorithm. Analyze the advantages and disadvantages of current ACO.Second: Analysis the reason of ACO's deficiencies. For deficiencies, proposed an improved ant colony algorithm. The new algorithm reference FACO to update the pheromone, reference the mind of Max Min Ant Colony(MMAS) to limit the size of pheromone of each side, reference the behavior of natural ant can't be perceived distance to change the choice of the path. Example shows, this improved ACO improve the performance of the algorithm in some extent. In particular, reduce the running time.Third: For the large number parameters of ACO and difficult to choose parameters. This paper use simulation method to determine the reasonable range of each parameters, discussing combinatorial optimization of five key parameters for ACO based on PSO, proposed combinatorial optimization program for ACO based on PSO. The program uses "three steps" strategies to determine the best combination.Four: On the basis of depth analysis of ant colony algorithm and it's parameters. Designing and implementation of ACO simulation system for TSP problem ,and the simulation system for the combinatorial optimization of ACO's parameters, which has strong practical significance.Five: Using parameters training system, Obtain the best combination of parameters, by using this improved ant colony algorithm, solving typical TSP problem, To verify effectiveness and feasibility of this algorithm and the best combination of parameters.In short, This improved ant colony algorithm is an effective method; Proposed scheme for combinatorial Optimization of ant colony algorithm based on PSO can overcome the limitations of traditional parameters selection. The design simulation training system for TSP parameters and parameters training system can promote the practicality of ant colony algorithm in some extent. Provide some reference for further research.