基于人工智能方法的无源光网络优化规划

摘要

使用人工智能方法对PON网络拓扑结构进行优化，以富士通FPX-1000无源光网络系统为例进行了仿真实验，并对遗传算法、增添算法和Hopfield网络方法进行比较，仿真结果表明遗传算法明显优于增添算法和Hopfield网络方法.遗传算法可使网络建设代价最小化，它非常好地解决了PON规划问题.%This paper uses artificial intelligence methods to optimize the PON topology. The simulation experiments of PON systems such as FPX-1000 are carried out. The genetic algorithm (GA)，add algorithm (AA) and Hopfield net approach (HNA) are compared,The simulation results show that GA is better than AA and HNA. GA can minimize installation cost of PON and realize the optimization of the passive optical networks with satisfaction.