Tactic Of Optimization Of Ant Colony Algorithm And Its Application

Ant Colony Algorithm is a new algorithm of heuristic method applied to optimization of combination in finding the solution or function optimization problems. It contains positive feedback and strong robustness, appropriately spread calculating system, which are easily combined with other methods, therefore, it has been extensively used to solve NP-hard problems. However, Ant Colony Algorithm still exist certain defects, namely, partial optimization, convergence and slow speed.Based on the defects in Ant Colony Algorithm, this dissertation presents three improvements on Ant Colony Algorithm searching tactics:(1) Make two variation points possible through Adaptive Mutated Variation and conduct a second ant colony search between the two variation points, reaching a new solution, through the adaptive self-variation in the initial solution, formed in basic Ant Colony Algorithm. Thus, possibility of local extremum can be reduced and optimizating ability and convergence speed in calculation can be improved. Based on what is mentioned above, the dissertation offers the Adaptive Mutated Ant Colony Algorithm (AMACA) .(2) After Brand and Bound Algorithm(BABA) is introduced into Adaptive Mutated Ant Colony Algorithm, the alternate node is optimized by the Brand and Bound Algorithm and inappropriate nodes have been removed, the Adaptive Mutated Ant Colony Algorithm can be conducted, therefore, a number of invalid searches can be eliminated and local extremum in calculation can be reduced, as a result, calculation optimizating ability and convergence speed can be improved. On this base, BABA-AMACA is raised.(3) After Artificial Fish School Algorithm(AFSA) is added to all iterations in Adaptive Mutated Ant Colony Algorithm, and the overall-situation fast convergence advantages in Artificial Fish School Algorithm are applied, convergence speed in Ant Colony Algorithm will be intensified. Food-searching act in Artificial Fish School Algorithm will improve the ability to help Ant Colony Algorithm break away from the local extremum. On this base, advanced hereby is AFSA-AMACA.The three improved Algorithms use to solve QoS routing optimization problems.It's calculation justifies the feasibility and efficiency.