Performance Comparison of Fuzzy-PSO and Fuzzy-GA for Double Girder Crane Control

Abstract

Double-girder overhead cranes are essential in industries for material handling, but their nonlinear dynamics, under actuation, and payload swing complicate control. This paper presents two Mamdani-type fuzzy logic controllers (FLC) using expert operator knowledge for trolley positioning and sway control. The crane’s nonlinearity was modelled via Euler-Lagrange and simulated in MATLAB/Simulink. Particle swarm optimization (PSO-FLC) and genetic algorithm (GA-FLC) tuned the FLC membership functions to minimize a composite cost function based on integral of time-squared error (ITSE). Simulation results showed PSO-FLC outperforms standalone FLC and GA-FLC. For position control, PSO-FLC achieves fastest settling time (3.8 sec), lowest overshoot (0.039%), and smallest error metrics. ITSE and ISE are reduced by 70% compared to standalone FLC, and by 43% (ITSE) and 40% (ISE) compared to GA-FLC. For sway control, PSO-FLC delivers 3.3 sec settling time, minimal overshoot (0.0049%), and negligible undershoot. IAE is reduced by 36% versus GA-FLC and 40% versus standalone FLC. Larger improvements in ITAE (44% and 49%) compared to IAE (36% and 40%) are significant, as ITAE penalizes steady-state error more heavily over time, indicating superior long-term error reduction and faster settling. Overall, PSO-FLC demonstrates the most effective strategy with superior precision, faster stabilization, and improved sway suppression for double-girder crane systems.