混合粒子群算法优化BPNN在模拟电路故障诊断中的应用

The Application of Genetic Algorithm Particle Swarm Optimization Algorithm Optimize the BPNN in Analog Circuit Fault Diagnosis

针对模拟电路的软故障,提出一种基于混合粒子群算法的BP网络方法来诊断模拟电路中的故障。该方法是把遗传算法和粒子群算法结合起来优化BP网络的权值和阈值,试图解决传统的BP网络在模拟电路故障诊断过程中易陷入局部最小的问题。详细阐述了该算法的实现,给出了该算法的详细流程图,并通过仿真实例比较了传统BP网络与混合粒子群算法优化下的BP网络在故障诊断中的表现,给出了实验实例仿真结果的图形和数据表格。由仿真图形和数据表格,形象直观地看出了两种算法运用在模拟电路故障诊断中的差别,验证了混合粒子群算法优化BP网络在模拟电路故障诊断中的有效性及可行性。

For the soft failure of the analog circuit, we propose Genetic Algorithm Particle Swarm Optimization algorithm based on the BP Neural Network （BPNN） method for fault diagnosis of analog circuit. This method combined Genetic Algorithm and Particle Swarm Optimization algorithm to optimize the synaptie weights and bias of BP Neural Network, trying to solve the BP Neural Network in analog circuit fault diagnosis process problem which easy falling into the local minimum. We elaborate every step of the algorithm in this thesis, then the detail flow char is given, and through a simulation example, we compared the performance in fault diagnosis between the tradi- tional BP Neural Network and the new method which discussesed in the paper called Genetic Algorithm and Particle Swarm Optimization algorithm optimize BP Neural Network, based on the results outputing in graphic or data forms, we can see the difference between the two algorithms application in fault diagnosis of analog circuit visually, draws a conclusion, demonstrate the validity and feasibility of the Genetic Algorithm Particle Swarm Optimization algorithm based on the BP Neural Network method applicationed in fauh diagnosis of analog circuit.