Genetic Algorithms-Based Methodologies for Design Optimization of Trusses

Abstract

Design optimization of trusses consists of arriving at optimum topology, configuration, and sizes of cross-sectional parameters, when weight of the truss is minimum, satisfying a set of specified constraints. Genetic algorithms-based methods provide ideal techniques for modeling many practical considerations, which make the optimal solution constructible. Two improved methods for size and configuration optimization and topology optimization are presented in this paper. The first one is a two-phase method, which in addition to arriving at better solutions compared to those obtained using earlier methods, also improves the computational efficiency. The second method, based on a variable string length genetic algorithm (VGA), addresses the topology optimization problem, taking into account a number of practical issues. Classical problems from literature are solved and the results are compared. Features of the proposed method, which help in modeling and application to optimal design of large truss structures, are demonstrated by solving a problem of a microwave antenna tower.