Author(s)

T. Jiang & P.Y. Papal ambros

Abstract

The method of moving asymptotes (MMA) represents a family of convex approximation methods suitable for structural optimization problems. Its efficiency depends strongly on asymptote and move limit locations. Second order information can be successfully employed for asymptote location but it is too expensive computationally. In this article, two strategies of approximating second order derivatives are derived and implemented in a new algorithm. A selection strategy for move limits to prevent oscillation in the presence of strongly coupled constraints is also described. Computational results for some typical problems are included. 1 Introduction An important consideration in algorithms for structural optimization is to reduce the number of structural analyses requi

Keywords