基于群代理的QoS-AP算法优化QoS组播路由

OPTIMISING MULTICAST ROUTING OF QOS WITH SWARMING AGENT-BASED QOS-AP

针对互联网组播应用中多约束服务质量(QoS)组播路由优化问题,提出一种基于群代理的融合蚁群(ACO)算法与粒子群优化(PSO)算法的QoS-AP算法。首先根据QoS约束,产生多个组播模型。然后利用ACO算法对每个模型和模型中的属性进行评估并放置信息素。再根据信息素值,利用PSO算法调整粒子代理的运动方式来重组组播树。经过多次迭代,最后形成一个满足QoS的最优组播树。通过仿真实验,与现有的PSOTREE、TGBACA算法进行比较。结果表明,该算法能够找出更好的组播树模型,不仅能够满足QoS约束,而且还最大限度地减少了树的成本。

For the optimisation issue of multicast routing of quality of service （QoS） with multi-constraint in Internet multicast applications, we propose a QoS-AP algorithm, which is based on swarming agents and integrates ant colony optimisation （ACO） and particle swarm optimisation （PSO）. First, according to QoS constraints it generates several multicast models. Secondly, it evaluates every model and the attribute of each model using ACO algorithm, and places the pheromones as well. Thirdly, according to the pheromone value, it reconstructs the muhieast tree by adjusting the motion ways of particle agent through pso. After several times of iteration, finally it forms an optimal multicast tree satisfying the QoS. Through simulation experiment, QoS-AP is compared with existing PSOTREE and TGBACA algorithms. Results show that, the algorithm is able to find better multicast trees model, and can meet the QoS constraints as well as minimise the cost of the tree.