Structural Design Optimization of Pallet Rack Structures Using Surrogate Methods and Metaheuristic Algorithms

Abstract In this study, a new framework for the structural design of pallet rack structures is proposed, utilizing a surrogate-based approach alongside metaheuristic algorithms. Initially, a simulation of a pallet rack structure was conducted based on laboratory data, employing a numerical finite-element model. This process established an explicit and surrogate equation using the response surface method to calculate the structure’s strength and volume. Subsequently, the performance of a meta-heuristic algorithm, referred to as the horse herd optimization algorithm (HOA), was enhanced and applied to optimize the dimensions of the pallet rack. The results indicated a 13% reduction in construction costs while preserving the stability and carrying capacity of the system within the allowable limits. Moreover, a comparative analysis of the proposed algorithm’s performance with other algorithms—namely, the whale optimization algorithm, genetic algorithm, grey wolf optimizer, moth–flame optimizer (MFO), and the original HOA—demonstrated its superior convergence rate and accuracy in the design of pallet rack structures. Additionally, an investigation of a three-story, three-span shelf frame, optimized based on the cross-sectional area determined by the proposed framework during the Bam earthquake scenario, revealed that the optimized structure, despite being 55 kg lighter, exhibited a 14.2% increase in energy absorption compared to the preexisting structure.