Modified Particle Swarm Optimization for Robust Anti-swing Gantry Crane Controller Tuning

摘要

This paper discusses an application of metaheuristic optimization algorithms for a single-objective constrained optimization in a robust feedback controller design of anti-swing gantry crane control. A set robust feedback controller gains is optimized based on plant's linear model having structured parametric uncertainty, i.e. gantry crane system. A wedge region is assigned as the optimization constraint to specify the desired closed-loop poles location which is directly related to desired time-domain response. The simulation results of the proposed robust control design using multiswarm particle swarm optimization without velocity (MPSOWV) is presented. The control performance of the feedback system optimized with MPSOWV are compared with that of PSO (particle swarm optimization), DE (differential evolution) and TLBO-PSO (improved teaching-learning-based optimization with the social character of particle swarm optimization). The simulation studies show that the controller optimized by the proposed MPSOWV demonstrates the most robust performance as compared to the other peer algorithms used in this paper for being able to produce the largest stability radius (r_c) consistently, i.e.r_c= 3.4325 in average for 20 runs.