基于分布式遗传算法以太无源光网络设计

针对以太无源光网络中各个光网络单元连接到光分配网络，再将光分配网络连接到光线路终端的全部通信费用为最小的问题，介绍了分布式遗传算法（DGA），该算法在分群操作的基础上，根据每个子群进化的优劣程度动态地划分子群大小、控制子群间个体的迁移方向，采用分布式遗传算法对以太无源光网络进行设计建模，利用分布式遗传算法使路径搜索更快，更好的收敛于最优的特点，使得光分配网连接到光线路终端的路径和为最小，解决了降低网络通信费用的问题。

Design on Ethernet Passive Optical Network Based on Distributed Genetic Algorithm

In order to connect every optical network unit with optical distribution network and connect optical distribution network with optical line termination, and minimize total communication cost, the ethemet passive optical network planning by using distributed genetic algorithm is optimized. Based on grouping operation, the algorithm dynamically allots the size of subgroup according to their performance, directs the migration. Compared to traditional genetic algorithm, simulation results show that distributed genetic algorithm makes the searching of routes converge to the global optimization faster and better.