Improved Genetic Algorithm In Multiobjective Optimization And Development Of Multiobjective Optimization Software

As one of the primary modern design methods, optimization plays a very important role in each field of industry. And most optimization problems are multiobjective optimization. Many multiobjective optimization problems involve simultaneous optimization of several competing objectives. Usually there is no single optimal solution, but rather a set of Pareto optimal points .In recent years, genetic algorithm has been widely used in finding non-dominated solutions. As to how to judge of the non-dominated solutions, there is no satisfied theory. In addition, no applied multiobjective optimization software is available.In this paper , the elementary theory and methods of multiobjective optimization and genetic algorithm is introduced. Some improvement of genetic algorithm is presented. Fuzzy evaluation is applied to parataxis choosing genetic algorithm to get better solution which satisfied both integral demand and individual demand of object functions. Besides, fuzzy decision is applied to Pareto-optimal solutions found by genetic algorithm with randomly specified weight values. In this way, weight value is unnecessary and best satisfied solution can be chose by fuzzy decision.A multiobjective optimization software is developed. It includes not only traditional methods but also modern methods basing on genetic algorithm. It is easy to use and is universal. It can promote the application of multiobjective optimization in some sense.