Evolutionary structural optimisation (ESO) for combined topology and size optimisation of discrete structures

Abstract

The single most researched and analysed topic in the field of structural optimisation in the last 40 years has been that of size optimisation for pin- and rigid-jointed frames. Much of the effort has been in developing the search engines that will find a global weight/volume minimum in the presence of multiple constraints, not all of which may be active at the minimum. Most of the algorithmic developments have followed formal mathematical programming techniques. Recently more heuristic methods have emerged for structural optimisation such as genetic algorithms, simulated biological growth and evolutionary algorithms. The authors have been researching the latter mentioned evolutionary structural optimisation (ESO) method for the last six years and have found it to be most efficacious for the full range of structural situations; topology, shape and size optimisation with stress, stiffness, frequency, stability constraints in 2D and 3D with single or multiple loads and support conditions. In the present work is a summary of some of our ESO research for pin and rigid-jointed 2D and 3D frames with or without multiple loads. The paper presents the simple evolutionary algorithm and gives examples covering the range of applications. Attention is also given to benchmarking the algorithms and to their reliability and robustness.