改进算子的遗传算法在QoS多播路由优化上的应用

针对QoS多播路由问题提出了一种改进遗传算子的遗传算法，通过设计多样性维持机制的选择算子、与代数相关的交叉算子和与个体适应度相关的变异算子，有效地解决了传统遗传算法应用于多播路由优化时的早熟收敛问题。实验结果显示，在随机网络中，该算法所获得的解与最优解的相对误差不超过1．0％，算法的收敛速度和寻优能力明显优于传统遗传算法。

Application of Genetic Algorithm with Improved Operators in QoS Multicast Routing Optimization

For QoS multicast routing problem, a genetic algorithm based on improved genetic operators is designed. By designing the selection operator with diversity maintaining mechanism, the crossover operator related to the number of generations and the mutation operator depending on fitness value, the algorithm has effectively solved the premature convergence problem in multicast routing optimization by using traditional genetic algorithms. The experiment results show that the relative error between the solution obtained by the algorithm and optimal solution is not more than 1.0%, its convergence rate and capacity of searching optimal solution are obviously better than traditional genetic algorithms in random networks.