Improved Particle Swarm Optimization Based on Entropy and Its Application in Implicit Generalized Predictive Control

Area of Science:

• Control Systems Engineering
• Optimization Algorithms
• Industrial Automation

Background:

• Industrial systems often face challenges with input-output constraints.
• Existing particle swarm optimization (PSO) algorithms can suffer from premature convergence and slow operation.

Purpose of the Study:

• To develop an improved particle swarm optimization (PSO) algorithm for implicit generalized predictive control (IGPC).
• To enhance the performance of IGPC in handling input-output constraints in industrial systems.

Main Methods:

• An improved PSO algorithm incorporating system entropy (SR) for a novel weight attenuation strategy and local jump-out mechanism.
• Modification of the velocity update mechanism and iterative adjustments to prevent local optimization.
• Integration of the improved PSO with gradient optimization for a rolling-horizon approach.

Main Results:

• The improved PSO algorithm effectively optimizes the performance index in predictive control.
• Simulation results demonstrate a reduction in system overshoot by approximately 7.5%.
• Settling time was reduced by approximately 6% compared to the standard PSO-IGPC.

Conclusions:

• The proposed improved PSO-IGPC algorithm offers superior performance in managing industrial system constraints.
• The enhanced optimization strategies significantly improve control accuracy and efficiency.
• This approach provides a robust solution for complex industrial control applications.