| With the rapid development of network communication technology, Mobile Ad HocNetworks(MANET)becomes one of the advantages of the computer industry hot topic,with its fast networking, flexible, and in military and civilian fields it has a broadapplication and development prospects.QoS Multicast Routing in Ad Hoc networkgenerates a minimal multicast tree while needs to meet the requirements of each constraintQoS and includes all multicast group members. Because QoS multicast routing problem isa nonlinear optimization problem, therefore, by referencing the ant colony algorithmsolves this problem,but the algorithm also exists some defects, such as too prematureconvergence and search capability slower and so on. Particle Swarm Optimization (PSO),an intelligent algorithm with global random optimization, it provides with simpleprograms, strong robust, efficient parallel computing and faster speed to global optimalsolution.Considering the impact of different ant colony algorithm control parameters onalgorithm performance,this paper improves ant colony algorithm, called dynamic antcolony optimization algorithm. Pheromone evaporation coefficient and pheromoneconcentration coefficient Q is the ant colony algorithm very important parameter, whichaffects the ability of ant algorithm optimization, parameter is a fixed value duringoperation algorithm; in dynamic ant colony optimization algorithm, based on changes inthe number of iterations, and Q that have a significant impact on the amount ofpheromone become into the dynamic change function, collaborating with each other antcolony algorithm parameters so that achieve the best results in the application ofoptimization and improve search performance. The improved algorithm on the TSP andthe Ad Hoc network multicast routing applications get more satisfactory results.Considering the ant colony algorithm own shortcomings, fuses with the particleswarm algorithms, introducing of a new multi-constraint functions for ant colonyalgorithm, so that the path of the ants find not only the lowest cost, but as much aspossible to meet the QoS constraints. The main strategy is to find a lot of convergencepath with the ant colony algorithm and selects the optimum alternative path set, then usesthe particle of particle swarm algorithm for further searching in the alternative pathsconcentration, by modifying the constraints and adjusting cross-operator gets the finaloptimal multicast tree and improves convergence in solving the problem of QoS multicast routing. |
