Application of Particle Swarm Algorithm to Optimization of PID Neural Network

摘要：

The original weight values of PID neural network are usually determined randomly and tend to sink into the local optimization in their learning process. To overcome the deficiency of PID neural network, the paper employs the particle swarm algorithm to optimize the PID neural network. To begin with, the particle swarm algorithm is used to acquire the optimized weight values of PID neural network. Next, by using the optimized weight values, we can optimize the PID neural network. Additionally, the performance of the improved PID neural network is assessed using a nonlinear coupling system. The simulation shows that the improved PID neural network effectively relieves the deficiency of the original PID neural network and has some obvious advantages in the calculation accuracy and convergence speed over the original PID neural network.

关键词： PID neural network particle swarm algorithm optimization

作者: Chi Yuan

作者单位: School of Mechanical and Electronic Engineering, Weifang University, Weifang, Shandong, China

会议类型: 国际会议

会议名称: 2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)

会议地点: 上海

会议语种:英文

页码: 582-584

在线出版日期: 2010-12-25（万方平台首次上网日期，不代表论文的发表时间）