A New Hybrid Differential Evolution with Gradient Search for Level Set Topology Optimization

Abstract

Topology optimization is an effective structural optimization concept for optimal design of engineering structures. However, it has many difficulties due to high number of design variables and complex problems same as compliant mechanisms and crashworthiness. Conventional methods for topology optimization does not have enough adaptability with current computer aided design (CAD) softwares and they are not powerful in solving difficult optimization problems. Level set which is a novel boundary tracking method had been recently used to solve problems in conventional methods. This paper is dedicated to propose a new hybrid method based on differential evolution (DE) and globally convergent method of moving asymptotes (GCMMA) to use both gradient direction of GCMMA and excellent exploration of DE. The method has been validated in familiar benchmark problems in compliance minimization.

Keywords

• Differential evolution; Method of moving asymptotes; Level set; Topology optimization.