摘要： 给出了基于分层图模型的光网络中链路阻塞率的计算公式，提出了一种改进的遗传算法，通过动态调整交叉和变异两算子，在全局搜索的情况下，调大变异算子的值，控制交叉算子的值；在收敛到局部进行搜索时，控制变异算子，提高交叉算子值，提高局部搜索能力，得到了一种快速的光网络阻塞率的计算方法，通过模拟显示：与普通遗传算法相比可以提高计算速度60%，并降低网络的阻塞率。

关键词: 光网络, 遗传算法, 动态变异交叉遗传算子, 阻塞率, NP-C问题

Abstract: The computing formula of the blocking probability for all-optical network with wavelength converters is presented. A quick dynamic genetic algorithm- variable mutation and crossover rates of genetic algorithm(VMCR-GA) is proposed to analyze the network blocking probability. The mutation and crossover operators in the VMCR-GA can be adjusted dynamically according to the search range to improve the speed of the algorithm. Comparing with the normal GA, the revised algorithm can improve speed 60% and reduce the network blocking by simulation on NSF network.

Key words: optical network, genetic algorithm, variable mutation and crossover rates of genetic algorithm, blocking probability, NP-C problem

中图分类号: 

• TP393