Topological optimization of a bumper beam via the messy genetic algorithm

Abstract

This paper introduces a technique for certain types of topological shape optimization problems, called variable primary and full grammars, which are derived from shape grammars. An industrial application is used as an example based on the optimization of an automotive structural bumper beam. The optimizer, the messy genetic algorithm, a technique more powerful than traditional genetic algorithms, does not use any derivative information on the evolution of the shape, but only simple evaluations. This technique produces interesting new designs for bumper beams, although they are not globally optimal. These new designs have moved from crenellations to more complex topologies.