Optimal Truss Sizing Using an Adaptive Differential Evolution Algorithm

Abstract

Truss sizing optimization is one of the structural design problems that are most difficult to solve, since it may have a nonconvex feasible region in cases of statically indeterminate trusses with stress and displacement constraints. The successful use of metaheuristics to solve such a design problem, which has been found to be effective, has been studied for several decades. This paper presents a new metaheuristic for truss sizing design. The method is based on differential evolution concepts while a strategically adaptive scheme is employed. Also a new, simple, but efficient constraint handling technique is proposed so as to effectively deal with design constraints. Numerical tests show that the proposed optimizer is powerful and can be compared with the best performers found in the literature.