Two-point gradient-based MMA (TGMMA) algorithm for topology optimization

Dual methods based on sequential approximations are usually employed for solving topology optimization problems. Among the approximation methods, the method of moving asymptotes (MMA) is perhaps one of the most popular methods used for solving these problems (Svanberg, 1987) [1]. However, recent investigations have shown poor performance of the MMA algorithm as compared to other approximations (Groenwold and Etman, 2010) [2]. In this paper we propose a two-point gradient based MMA approximation, termed as TGMMA, to improve the performance of the MMA algorithm. Numerical results demonstrate the efficiency of the TGMMA algorithm, which improves the MMA algorithm and also shows better performance over other existing approximations.