A Hybrid Genetic Algorithm for Structural Optimization Problems

This paper presents a hybrid optimization algorithm based on a modified genetic algorithm (GA). The algorithm includes two stages. In the first stage, a global search is carried out over the design search space using a modified GA. In the second stage, a local search is executed that is based on GA solution using a discretized form of Hooke and Jeeves method. The modifications on basic GA includes dynamically changing the population size throughout the GA process, utilizing variable penalty multiplier and the use of a square root form of the penalty function in constraint handling. The hybrid algorithm and the modifications to the basic GA are examined on the design optimization of a well-known test problem (10 bar truss). The effect of different parameters and techniques of handling GA operators on the performance of the proposed algorithm is investigated. The hybrid algorithm is employed for the design optimization of a reinforced concrete flat slab building and the results are compared with those of using the GA only.