QoS组播路由的多种群遗传算法

Study on QoS Multicast Routing Algorithm Based on Multi-Population Genetic Algorithm

针对单种群遗传算法在求解QoS组播路由问题中存在的容易早熟、收敛性差等缺陷,提出了一种基于多种群遗传算法的QoS组播路由算法。该算法在初始化过程中采用多种初始化算法生成了不同种群,并设计了多种交叉、变异操作,保证了种群间进化过程的独立性和算法的多样性;增加了种群间协调机制,提高了算法的整体收敛性。仿真结果证明了多种群遗传算法的有效性和优越性。

In order to solve the anti-precocity and convergence of the single-population genetic algorithm when applied to multiple QoS constrained multieast routing problem, a method of multi-population genetic algorithm based on cluster route is proposed. To prove evolutionary independence among the population and the variety,various initial algorithms is used to generate multi-population, and different cross-operation and mutate-operation are also designed. To improve the whole convergence ,an correspond-policy is appended among the populations. The simulation results prove the validity and su- periority of the multi-population genetic algorithm.