Large-scale Combinatorial Optimization Problems Ant Colony Algorithm Applied Research

Ant colony algorithm (ACA) is a new kind of simulated evolutionary algorithm. It is proposed by Italian Macro Dorigo, Ant colony algorithm has many good effects when it has solved the combination optimization problem. However, when it comes large scale combinational optimization problem, its searching speed is slow. In order to successfully solve large scale combinational optimization problem, in this paper, on the basic of deeply study of large scale combinational optimization and Ant colony algorithm, I brought up a strategy and algorithm for solving large scale combinational optimization problem.Being of basic ACA is easily combined with other algorithm, such as clustering algorithm, Genetic algorithm and data mining, I brought up four measures. Firstly, a new parameter evolution algorithm of ant colony, an evolution algorithm about parameter of ant system, is proposed in this paper here. Due to this evolution algorithm, I found perfect configuration of parameter, and this parameter can improve the speed of running of ACA. Secondly, large scale combinational optimization problem was provided some son problem of little scale problem by using clustering algorithm, at the same time, aim at wave of clustering algorithm, some discuss about it was brought up. Thirdly, by using of data mining algorithm, I found out some local optimization path, by merger of path of local optimization, the scale of problem was reduced, at the same time, the time of running of ACA was reduced, Lastly, information entropy was introduced in this paper, information entropy instead of iterative degree was used of ending condition.The result of the experiment suggests that the improved algorithm is effective.