神经网络的设计主要集中在网络权值的选取和神经网络结构的确定两个方面 ,与遗传算法结合是目前研究发展的趋势 .本文与一般的基于遗传算法的神经网络设计相比 ,提出一个新型算子—— BP算子 ,并对神经网络的权值和结构同时优化 .仿真结果表明该算法结果比较理想 . The design of neural networks is mainly focused on the selection of network weights and the determination of neural network structure, and the combination with genetic algorithms is the trend of research and development.This paper proposes a new type of neural network compared with the general design of neural networks based on genetic algorithms Operator - BP operator, and simultaneously optimize the weight and structure of the neural network.The simulation results show that the algorithm is ideal.