Neural Network Intelligent Control Based on MPSO

Abstract

With the increasing complexity of mechanical equipment, the control effectiveness of traditional intelligent control systems can no longer meet the needs of modern industrial production. In order to reduce errors in intelligent control systems while ensuring system performance, this study proposes a new Particle Swarm Optimization (PSO) optimization scheme. The study simplified the PSO algorithm from three aspects: algorithm parameters, speed, and position formula, and corrected the formulas for individual optimal values and global optimal values. Research will name the optimized algorithm Modified PSO (MPSO). On the basis of the MPSO algorithm, neural network intelligent control has been innovatively improved. In the experimental results, the MPSO optimized controller controlled the error within 0.01 within 0.02 seconds. At this time, the Whale Optimization Algorithm (WOA) optimized error was 0.072, and the PSO optimized error was 0.478. Compared to PSO and WOA, the control error of MPSO has decreased by 98.95% and 93.06%, respectively. In addition, the proposed method not only has the best control effect, but also has the shortest system response time, with an average time of 1.294 seconds. Compared to PSO and WOA optimization, it reduces by 61.48% and 43.07%, respectively.The results verified that the proposed method in this study can effectively improve the accuracy of intelligent control and control the error within the target range within 0.02 seconds. The research not only simplifies the calculation of the PSO algorithm, but also effectively reduces the error of the algorithm, providing a reference for research in the field of intelligent control.

Keywords

• Intelligent control
• nonlinear problems
• neural network
• particle swarm optimization (PSO)