Ant Colony Algorithm

Ant Colony Optimization is proposed in 1991 by Italian researchers Dorigo M., Maniezzo V. and Colorni A. It is a newly bionic optimization algorithm which can imitate ant colony intelligence action in insects societies. With the extraordinary merits of distributional concurrency, stronger robustness and easily combinating with other algorithms, this algorithm attracts more and more domestic and foreign scholar's attentions, and becomes the research focus and front question of heuristic algorithm at present in domestic and foreign.In view of fact that strict theoretical foundation has not yet been built, domestic and foreign related researches are in experimental exploration and initial application stage, however, the research of ant colony algorithm has penetrated from originally single Travelling Saleman Problem field to multi-application fields, from solving one dimensional static optimization problems to solving multi-dimensional dynamic combinatorial optimization problems , and from discrete domain to continual domain. Thus, this newly rising bionic optimization algorithm shows vigorous vitality and wider development prospects.At the foundation of reading lots of domestic and foreign literature material , the article summarizes currently research productions and problems which would urgently be solved, and introduces the development of algorithm, the basic theory of algorithm, kernel ideal, mathematics model , the conception of Ant Colony Optimization and implementing framework, implementing framework and process of basic Ant System algorithm, and a serial of proved algorithms which were based on basic Ant System algorithm. At the sametime, having programming and debuging in matlab, basic Ant System algorithm and Ant Colony System algorithm are achievedIn light of parameters configuration has not been perfect theory basis, algorithm has been more dependent on parameters, and parameters, such asα,β,ρthere select method and principle have directly related to convergency and optimization capacity of algorithm, therefore, the article carries out a mass of parameters configuration tests for better test results, and detailed parameters analysis, as a result, makes certain the combination of parameters configuration. At the basis of completion of basic Ant System algorithm and Ant Colony System algorithms'parameters analysis, this article attempts two algorithm improvements, and brings forward two improvement schemes, which add new improvement strategy to initial pheromone matrix and second search by clustering in local search part. At last, this article finishes basic Ant System algorithm and improved Ant Colony System algorithm, and takes Olive 30 of TSPLIB problems as a test example. The result of imitated optimization shows that the improved algorithm has better convergency speed and ability of global optimization than the corresponding period results.