A Hybrid Ant Colony Algorithm For Solving TSP

Ant colony algorithm (ACO) is a heuristic intelligent algorithm which i s widely used in solving combinatorial optimization problems.It is shows a superior performance and potential in solving complex discrete optimizatio n problems and has a good application in many field. But there are still som e shortcomings, such as spending a long time in searching, easy to fall in to the phenomenon of precocity and stagnation. Ant colony algorithm also ha s a characteristics that easy to integrate with other algorithms, while there ar e many other new bionicalgorithm have a very good performance in solving combinatorial optimization problems and complement each other’s advanta ges to ant colony algorithm.Shuffled frog leaping algorithm (SFLA) is a newly swarm intelligence optimization that combined with the characteristics of Memetic algorithm of genetically transmitted characteristics and Particle Swarm Optimization.Simulated Annealing Algorithm(SA) is a random optimization algorith m from the principle of solid annealing,it is effective in solving the Large-scale TSP problem,is one of important research fields in recent years. And consequently, based on the background, the main research content s in this paper are as fallows:First of all, the basic theory of ACO, SFLA and SA are reviewed by combining with the TSP problem, including the research and developmen t of ACO, the mechanism of the algorithm, the algorithm model and algorithm implementation.Secondly, some improved strategies are suggested according to the disa dvantages of ACO, including routing and pheromone updating method, thu s further improve the calculation precision of the algorithm.Third, SFLA and SA are separately combined with ACO.Some improv ed strategies of ACO and hybrid strategies of the hybrid algorithm are pr oposed base on SFLA and SA. Including the initialization of pheromone, dynamic engagement strategy in SA-ACO and SFLA-ACO, the parameter optimization base on the the Gauss mutation operator and Cauchy mutatio n operator in SFLA-ACO and improvement of cooling formula in SA-AC O.At last, the improved algorithms are validated by experiments, the vali dity of the algorithms are proved by the experimental results.