Feasibility Rules Assisted Gravitational Search Algorithm for Optimization of Design Engineering Problems

Abstract

Metaheuristics are advanced optimization techniques used in various real-world applications for achieving the best value for a particular cost function. The search landscapes of most real-world problems are restricted due to constraints of both equality and inequality types. Since metaheuristic approaches are generalized in nature, assisting them with a constraint-handling mechanism can optimize the constrained problems. The gravitational search algorithm (GSA) is a well-known physics-based algorithm used to optimize several benchmark and real-world problems. The present study assists GSA with feasibility rules to find the optimum solution in a constrained landscape. To obtain enhanced performance than GSA, feasibility rules are coupled with chaotic GSA (CGSA). Ten constrained problems related to mechanical, civil, and chemical engineering are solved. Mean and standard deviation values justify that feasibility rules with CGSA give better results. A statistical test (Wilcoxon’s signed rank test) confirms the superior performance of CGSA.