A Two-stage Optimization Algorithm Of NSGA2 Guided By The Improved Bee Evolutionary Genetic Algorithm And Its Application

The research issues of modern engineering are increasingly complex, and they often have multiple inter-constraint design goals, which make the theory and method of the traditional multi-objective optimization problems continue to be challenged. As the main method to solve this kind of problem s, the multi-objective optimization algorithms should have the characteristics of strong robustness, high efficiency and high accuracy. In recent years, many excellent multi-objective optimization algorithms and theories have been proposed. In the paper, the advantages and disadvantages of various optimization algorithms are deeply studied through the analysis of the research status of the optimization algorithm s. For the shortages of classic NSGA2 in solving complex problems, combined with the advantages of bee evolutionary genetic algorithm, a multi-objective evolutionary algorithm called NSGA2 guided by the improved bee evolutionary genetic algorithm is designed to execute two-stage optimization, and the application of the proposed algorithm is studied. The main work and innovation points are summarized as follows:1) By optimizing the selection operator ?, according the similarity of cross parents to select cross methods and according the performance of individual's parent to select mutation methods, the bee evolutionary genetic algorithm and the searching efficiency of the first phase are improved. Then, through the analysis of the shortages of classical NSGA2 in solving complex and demanding optimization problems, based on the improved cross-way and mutation methods, this paper proposes to delete duplicate individuals, introduce new individuals, use elite population instead of the parent population and other improvement measures to improve the population diversity, algorithm efficiency and accuracy in the second phase. For the optimization problems with demanding constraints, this paper introduces an external auxiliary population. The infeasible solutions which the constraint violation degree are smaller than a given value are copied to the external set and participate in the next generation of evolution to improve the population diversity and guide other infeasible solutions rapidly approaching the feasible solution boundary. Based on the above improvements and the thought of the two stage optimization algorithm, the corresponding detailed algorithm flows are designed to respectively solve the multi-objective problem with demanding constrained conditions and the multi-objective problem with no constraint.2) From the point of green performance, a VRPSPD model of recycling and remanufacturing of waste products is established, and the proposed algorithm is used to solve the optimization model. Compared the results of different customer size calculated by the proposed algorithm with the results calculated by the algorithm improved before, the efficiency of the proposed algorithm is verified.3) Use the proposed algorithm to solve the multi-objective flexible job shop scheduling problem with the objectives of minimizing the maximum completion time, the maximum machine load and the total load of the machine s. The validity of the proposed algorithm is illustrated by the calculation of the basic examples and the comparison of the results calculated by other six classic algorithms.