Multiple Optimum Size/Shape/Topology Designs for Skeletal Structures Using a Genetic Algorithm

Abstract

A genetic algorithm is presented that can simultaneously optimize size, shape, and topology of skeletal structures, including both trusses and frames. The algorithm is unique because it finds multiple optimum and near-optimum topologies in a single run. The algorithm was executed on a bridge example where it found both traditionally recognized bridge topologies as well as less familiar topologies. It was also executed on two standard test problems as well as on a plane frame example. This algorithm presents the designer with more choices and more information than algorithms that converge to a single optimum design.