Nonlinear Model Predictive Controller for Maglev Systems : A Support Vector Machine Approach Optimized with Particle Swarm Optimization

Abstract

Model Predictive Control is employed by the plant's linear model, which incorporates a multilayer feed-forward network for predictive control. The inclusion of Particle Swarm Optimization as the optimization technique results in a significant decrease in the required convergence iterations. This paper presents a new approach that uses Particle Swarm Optimization as the reduction technique in conjunction with SVM-generalized predictive control. This study employs a mathematical model based on a Support Vector Machine, with the objective function optimization model being Particle Swarm Optimization. The strategy utilizing SVM exhibits a high level of accuracy, as the controller's precision is contingent upon the quality of the model. Simulation findings show that an acceptable solution can be reached in just two iterations. empirical proof also supports real-time control's feasibility.