基于混合遗传算法的神经网络在智能故障诊断中的应用

设计了用模拟退火的混合遗传算法代替BP网络的反向传播过程的改进算法，解决了在故障诊断系统中BP算法容易陷入局部极小值的问题．该算法是在遗传算法中引入模拟退火机制，将其同BP算法结合，形成一个混合的优化算法．新算法既有神经网络的学习能力和鲁棒性，又有遗传算法的强的全局随机搜索能力．仿真结果表明，这种改进算法极大提高了内燃机故障诊断系统的效率和准确性．

Application of Neural Network Based on the Mixed Genetic Algorithm in Failure Diagnosis

A mixed genetic algoritbm with simulated annealing is designed for replacing the backward propagation process algorithm of BP network to solve the problem that BP algorithm is prone to get the local minimum in the failure diagnosis system. Through introducing the simulated annealing mechanism into the genetic algorithm and combining it with BP algorithm, a mixed optimized algorithm is formed. The new algorithm is of both the learning ability and robustness of the neural network,as well as the strong global random searching ability of the genetic algorithm.The simulation results indicate that the improved algorithm can greatly improve efficiency and accuracy of internal combustion engine failure diagnosis system.