Research In Multi-object Optimization Genetic Algorithm

In view of importance of multi-object optimization in engineering, economy, management, military and so on, the research on multi-object optimization has been paid more attention, it has developed into a new branch of science and demonstrated powerful vitality in application. The genetic algorithm is a global optimization, auto-adapted, probability search algorithm which uses the experience of biological natural selection and genetic mechanism for reference, owing to its unique superiority and robustness in solving the complex system optimization, it becomes a very effective method in solving multi-object optimization problems.This paper dwells on present situation of the research, basic principles, representative algorithm about the multi-object optimization, it also presents the mathematical theory and implementation technology of genetic algorithm, it studies some problems which have not been solved in multi-object optimization genetic algorithm. Through research of elitist strategy, the paper presents a multi-object optimization genetic algorithm based on data warehouse of Pareto optimal solutions, the algorithm preserves the individuals of Pareto optimal solutions from each generation in the data warehouse, it eliminates the same or similar individuals in the data warehouse by measuring the distance between two individuals , not only the new algorithm enhanced the performance of algorithm, but also it improves the quality of solutions, it can obtain the massive and well-distributed Pareto optimal solutions. In view of the constraint condition in the application of multi-objective optimization problems, the paper presents a multi-object optimization genetic algorithm with complicate constraints based on population classification. This algorithm lays special stress on population diversity, and solves it using K-means clustering analysis, it divides the population into four classes, the different population are given the different fitness, so it can embody elitist strategy. A large number of calculations using the computer indicate that not only the algorithm can get the massive and well-distributed Pareto optimal solutions, but also the evolution rate is extremely quick, it can obtain extensive Pareto optimal solutions easily only 10-40 generations or so.