Structural design employing a sequential approximation optimization approach

Abstract

This paper presents an improved sequential approximation optimization (SAO) algorithm that is suitable for structural design optimization tasks. First, an adaptive sampling strategy is proposed to balance between the competence to locate the global optimum and the computation efficiency in the optimization process. Furthermore, an original estimation of the width of the basis function is proposed based on the local density of sampling points, which enhances the RBF for the SAO. The efficacy of the enhanced SAO algorithm is validated using several benchmark structural design cases and the computing costs are substantially reduced in comparison to other optimization algorithms.