Research Of Ant Colony Algorithm & Its Application On Network Routing Optimization

Ant colony algorithm is a novel category of bionic meta-heuristic system,and positive feedback with parallel autocatalytic mechanism are adopted in this algorithm.The ant colony algorithm,which has strong robusticity,excellent distributed computing mechanism and is easy to combine with other methods in optimization,has shown an tremendous development potential in various combined optimization fields for its outstanding performance.The research question has absorbed lots of scholars at home and abroad.It becomes a international research focus and front-subject in Computational Intelligence field nowadays.In this paper,first we investigate the research status of ant colony algorithm in and abroad, analyse the major merits and shortcomings of it, study the principle and mathematical model of ant colony algorithm in depth,and bring in two improved ant colony algorithms of ACS algorithm(ant colony system) & MMAS algorithm(Max-Min Ant System).Then,aiming at the flaws of ant colony algorithm,QMACO algorithm based on traveling salesman problem(TSP) is proposed to optimize the system.The improvements are mainly on 3 aspects:I,QPSO algorithm idea is brought in to solve the problems at prophase,such as blindfold ant searching and slow convergence speed. First we make use of the rapidity and global convergence of QPSO algorithm to carry on early stage searching,then the located each particale's historical optimum value is regarded as the initial distribution of pheromone in later stage.Thus the blindfold ant searching at prophase can be avoided.II,In order to overcome the drawback of being trapped in local optimum,we introduce multi-behaved ant,fully utilize the positive feedback mechanism and high solution efficiency to obtain the optimum value.III,Improve ant colony algorithm by synthesizing the merits of both ACS algorithm and MMAS algorithm.Also simulation experiment is done toTSP,the results demonstrate that QMACO algorithm have strong optimization ability in solving TSP.Followed that,we apply QMACO algorithm to network routing problem, work on algorithm design on unicast routing and multicast routing problem respectively, finally proposed different models of algorithm. The simulation results show that QMACO algorithm is reliable and efficient in solving QoS routing problem.The last is a summing-up and we put forward some suggestions for the future studies.