Applicability of Subspace Harmony Search Hybrid with Improved Deb Rule in Optimizing Trusses

Abstract

This study addresses the optimization of truss structures with discrete design variables using a subspace harmony search (SHS) algorithm integrated with an improved Deb rule. Harmony improvisation and updating of the SHS are implemented in a subset randomly chosen from harmony memory (HM) to avoid premature convergence. The SHS also introduces a local search strategy with a dynamically activated possibility to enhance exploitation ability. The proposed improved Deb rule can filter redundant constraint violation evaluations during the optimization process and greatly reduce the required number of analyses in structural optimization in comparison with the conventional approaches. The numerical investigation, including four truss structure weight minimization problems, compares the search ability and computational efficiency of the proposed approach with other metaheuristic algorithms. The numerical results illustrate the robust search ability of the proposed SHS. The improved Deb rule leads to significant reduction in computational cost without compromising either the search ability of HS-based approaches or the efficiency of the original Deb rule.