A Reformulation of the Ant Colony Optimization Algorithm for Large Scale Structural Optimization

摘要

This study intends to improve performance of ant colony optimization (ACO)rnmethod for structural optimization problems particularly with many design variablesrnor when design variables are chosen from large discrete sets. The algorithmrndeveloped with ACO method employs the so-called pheromone scaling approach tornovercome entrapment of the search in a poor local optimum and thus to recoverrnefficiency of the method for large-scale optimization problems. Besides, a newrnformulation is proposed for the local update parameter in the algorithm. The efficacyrnof the proposed algorithm is quantified using two numerical design examples chosenrnfrom practical size optimum design of steel structures. The results obtained with thernproposed algorithm are compared with those of other methods, such as particlernswarm optimization (PSO), harmony search optimization (HSO) and geneticrnalgorithms (GAs). The design problems are formulated according to the provisionsrnof ASD-AISC (Allowable Stress Design Code of American Institute of SteelrnInstitution).