Portfolio Research Based On Improved Genetic Algorithm

With the great improvement of China's economic strength, a lot of investment projects have strong development potential. In the face of numerous investment instruments, how the investors choose the optimal portfolio investment is more and more significant. Markowitz first put forward about portfolio Mean-Variance model in 1952.Mean-Variance Model is the base of modern portfolio theory. It is the beginning of financial investment quantitative research. As the major topic of financial investment theoretical research and an important tool for decision-making practice, core foundation is composed of it in modern portfolio theory. However, in real economic life, this model is limited. Genetic algorithm has a prosperous development period in the 90's. It becomes very popular topic both in theoretical research and applied research. Many scholars began to use genetic algorithms to solve different portfolio model. However, it has been more difficult in solving improved multi-objective M-V model.This paper first introduces the typical genetic algorithm idea and the steps. It proposes an improved genetic algorithm Objective Adaptive Parallel Genetic Algorithm based on analyzing the genetic algorithm performance bottlenecks. Second, it applies Objective Adaptive Parallel Genetic Algorithm to solve the Markowitz model which is multi-objective limited investment restrictions. In this process, it discusses operator parameter design and studies dynamic adjustment specie size and specie diversity on the impact of the crossover and mutation probability technology. Matlab environment is used to compile programming for solving the model and simulating genetic algorithm search process. Research results show that improved genetic algorithm effectively improves the efficiency of the algorithm. This method is scientific and reasonable. To a large extent, it is in line with actual conditions in China.