A Hybrid Particle Swarm Optimization and Genetic Algorithm for Truss Structures with Discrete Variables

Abstract

A new hybrid algorithm of Particle Swarm Optimization and Genetic Algorithm (PSOGA) is presented to get the optimum design of truss structures with discrete design variables. The objective function chosen in this paper is the total weight of the truss structure, which depends on upper and lower bounds in the form of stress and displacement limits. The Particle Swarm Optimization basically modeled the social behavior of birds on the basis of the fact that Individual birds exchange information about their position, velocity, fitness, and on the basis that the behavior of the flock is then influenced to increase the probability of migration to other regions with high fitness. One of the problems of PSO is that it is easily trapped at the local point due to its non-uniform movement. The present study uses the mutation, random selection, and reproduction to reach the best genetic algorithm with the operators of natural genetics. Therefore, only identical chromosomes or particles can be converged. In other words, PSO and GA algorithm goes from one point in the search space to another point, interacting with each other. In this way, this helps them to find the optimum design by means of deterministic and probabilistic rules. The present study merged the two algorithms together in order to design several benchmark truss structures, and then the results of the new algorithm compared to those of other evolutionary optimization methods.

Keywords

• particle swarm optimization
• genetic algorithm
• size optimization
• structural optimization
• discrete variables