| The goal of this research is to develop a constrained multiobjective optimization method in the form of a robust, practical, problem-independent algorithm, and to investigate the effect of multiobjective optimization on structural design.; This research studies the properties and characteristics of multiobjective optimization and solutions. Design objectives, constraints and design variables as well as their effect on structural design and behaviors are investigated. Dynamic input energy for a multidegree-of-freedom structure is formulated. A multiobjective optimization algorithm based on game theory is developed. Fuzzy set theory is introduced to construct a fuzzy constrained environment. A Pareto genetic algorithm is developed based on a simple genetic algorithm. Two new genetic algorithm operators--multiobjective fitness function and niche method--are proposed. A new genetic algorithm process--Pareto set filter--is added to genetic algorithms. Finally, a multiobjective optimization algorithm, which is robust and performs uniformly well on a broad range of problems, is developed with genetic algorithm, fuzzy logic and game theory.; Six frame structures and five truss structures have been optimally designed with static load, seismic excitation, and active control. Materials for the structures are steel, reinforced concrete, and a combination of both. Optimization goals cover weight, structural cost, displacement, strain energy, input energy, potential energy, and performance index. Constraints include stress, displacement, frequency, and ratio of story stiffness. The numerical results show that the newly developed algorithms can locate a global solution and Pareto optimum set which are usually difficult to find. |
