Research On Application Of The Portfolio Problem Of Internet Fund Products With Multi-Objective Evolutionary Algorithm

In recent years, the Internet financial money-management has been more and more popular and gotten much attention with the rapid development of Internet technologies, such as mobile Internet, cloud computing and so on. A variety of fund financial products can be easily purchased through the online platform and mobile phone applications. There are a lot of fund financial products on the market currently. How to select the optimal portfolio in these fund products is a very important and significant problem for individual investors and the Internet financial platform.How to select the optimal portfolio in these fund products is a kind of portfolio problems in essence. This is a typical multi-objective optimization problem. It mainly focuses on two important goals, which are the Risk and the Return. In practice, it is difficult to solve this problem, because it needs large calculations when we consider many kinds of funds and many restrictions such as tansaction cost, the number of funds and so on. The multi-objective evolutionary algorithm is very effective to solve the complex optimization problem and is widely applied to solve the portfolio problem. This paper puts forward the improved model based on the classical model of portfolio investment and improves the initial data processing method, coding and genetic operators of the multi-objective evolutionary algorithm according to the actual problem. The experiments show the improved algorithm significantly improves the computational efficiency and the effect of application.The main work and innovation of this paper are as follows:1. Establishe an improved multi-objective optimization model based on the classical mean-variance model, which considers the transaction costs, the number of funds and purchase price;2. Study the typical multi-objective evolutionary algorithm——NSGA-II, realize it for solving the model, and compare its convergence and time performance with other multi-objective evolutionary algorithms;3. Propose an improved algorithm based on NSGA-II, which improves the processing mothod of the initial data by using a nondominated sorting to reduce the decision space, uses the imporved binary code to take place of real coding with the integration of constraint conditions of the improved model to reduce the model’s complexity. These measures significantly improve the computational efficiency and effectiveness.4. To verify the effect of the improved algorithm, this paper adopts two data sets, including the standard data set and the actual fund data set, which is crawled from the Internet. Otherwises, this paper have done lots of contrast experiments and analyses.In this paper, we make a detailed study of the portfolio problem of fund, set up a practical and effective multi-objective optimization model, and realized the typical multi-objective evolutionary algorithms for the optimal solution of the model. Besides, in order to solve the problem of low efficiency and poor application effect of algorithms, this paper has put forward the improved method. Therefore, this paper has a certain theoretical and practical significance to the application and development of the multi-objective evolutionary algorithms and the selecting of the best portfolio for individual investors and fund platform.