Bar‐system representation for topology optimization using genetic algorithms

Abstract

PurposeThis paper proposes a bar‐system graph representation for structural topology optimization using a genetic algorithm (GA).Design/methodology/approachBased on graph theory, a graph is first used to represent a skeletal structure consisting of joining paths in the design domain, each of which is represented by a chain subgraph with finite number of vertices. Based on the edges of this graph, a bar‐system representation is proposed to define all the bars and the resulting topology is obtained by mapping each bar with its corresponding thickness to the design domain which is discretized into a regular mesh. The design variables are thus reduced to the spatial distribution of the vertices and the thickness of each bar. This method combines the advantages of both continuum and ground structure optimization methods.FindingsThe overall procedure is applied to classical structural topology optimization problems and its good performance is illustrated in the numerical examples.Originality/valueIt is suggested that the present representation method is both physically meaningful and computationally effective in the framework of topological optimum design using GAs.