Optimized Design of Two-Dimensional Structures Using a Genetic Algorithm

Abstract

A design procedure incorporating a simple genetic algorithm (GA) is developed for discrete optimization of two-dimensional structures. The objective function considered is the total weight (or cost) of the structure. The objective function is minimized subjected to serviceability and strength requirements. The GA-based design procedure FEAPGEN is developed as a module in the Finite Element Analysis Program (FEAP). Special features of FEAPGEN include discrete design variables, an open format for prescribing constraints, design checking using the American Institute of Steel Construction Allowable Stress Design (AISC-ASD) specifications, multiple loading conditions, and a comprehensive AISC database of available structural steel members. Several strategies for reproduction and crossover are investigated. In particular, a group selection scheme for reproduction that does not require fitness scaling is applied. Various fitness and penalty functions are investigated for their appropriateness to the ASD design of two-dimensional structures. A comparison is presented between FEAPGEN genetic search design procedure and a classical continuous optimization method based on the optimality criterion.