Research Of Elitist NSGA And Its Application In Regional Water Resources Optimal Allocation

Multi-objective optimizations abound in science research and engineering, classical multi-objective optimization methods have several shortcomings in high dimension and multi-modal problems. With the progress of the research of evolution algorithms, domestic and overseas researchers have developed many multi-objective optimization genetic algorithms based on Simple Genetic Algorithm, then proved that the Non-dominated Sort Genetic Algorithm(NSGA) shows the great advantages among them, later, some researchers integrated the elitist strategy with NSGA, then proposed the improved algorithm as Elitist Non-dominated Sorting Genetic Algorithm(NSGA-â…¡).For the research of Elitist Non-dominated Sorting Genetic Algorithm, both theoretical and practical perspectives are conducted extensively in overseas countries, but in our country it is rather limited. In this paper, first study and do some research on the basic theory, then give an improved algorithm proved by experiment, at last employ the effect into application. The main contents are as follows:(1) Introduce the current state of the research on multi-objective optimization and Multi-objective Optimization Genetic Algorithm, mainly about the Non-dominated Sorting Genetic Algorithm. Start with the concept and the theoretic of multi-objective optimization, dissertate the basis of principle, flow of the algorithm and the essence of strategies for both NSGA and NSGA-â…¡,then bring forward some shortages of current algorithm.(2) Focusing on the problems that the searching abilities of Non-dominated Sorting Genetic Algorithm powerful in global but weak in local and lack of using of the info correlate with the evolution, by using the ideology of Hybrid Multi-objective Evolutionary Algorithm for reference, integrate the Tabu Search as local search algorithm with NSGA-â…¡, propose a algorithm of NSGA with TS local search and elitist strategy. First modify the Tabu Search Algorithm for adapting multi-objective optimization, for using the info correlate with the evolution, propose a self-adapted neighborhood searching strategy, in addition, construct a new crossover operator for enhancing the searching ability. At last, use the Eckart Zitzler test function for testing, constitute 4 targets, then analyze the testing result.(3) Regional Water Resources Optimal Allocation is a typical multi-objective optimization matter, this paper first analyze the mathematic model, then employ the theories into practice, set up a new method for solving the problem. Using the regional water resources optimal allocation of Xiamen, compare with Large System Overall Optimization Genetic Algorithm, it shows that the method bring forward by this paper the superiority over others, it can satisfy the requirement of creating some decision-making solutions, all indicate that this research have some practical significance.