Study On Artificial Fish-Swarm Algorithm And Its Application In Multi-objective Portfolio Problem

In today’s all walks of life as well as people’s daily activities, any form of investors, businesses or individuals, they are often faced with the opportunity to portfolio investment. In the investment process, investors will be considering various factors, which refers to a multi-objective portfolio problem. In the current period of rapid development of the market economy, this problem has already become one of the hot research topics in the field of financial investment.The description of the problem is that from each of investment, which has a specific income, transaction costs, risks and other characteristics, investors choose the appropriate number of assets, and determine their best investment ratios in the total investment decision, in order to obtain a better portfolio decision, at the same time to ensure that investment risk is minimized, while transaction costs are minimized, then ultimately achieve investment returns’maximum. Against this kind of multiple targets, that exists conflicting nature, how to make the best investment decisions, this issue has been one top topic of the financial sector and academic sector.Multi-objective portfolio problem is a relatively complex problem in the field of operations research, and it belongs to NP-hard problems. There is a certain degree of difficulty when using traditional optimization methods to solve the problem. In recent years, many scholars have studied using intelligent optimization algorithm to solve the multi-objective portfolio problem, and have done a lot of attempts and in-depth exploration, this has become a more popular research topic, and have achieved good results. But so far, few scholars uses Artificial Fish-Swarm Algorithm to solve multi-objective portfolio problem, therefore, this paper proposes using Artificial Fish-Swarm Algorithm to deal with the problem. The full paper presents and discusses about application of AFSA in multi-objective portfolio problem, and proposes two improved AFSA, which counter several deficiencies of AFSA, then verifies feasibility of the basic AFSA and improved algorithm through testing functions. At last, this paper introduces to use MATLAB to simulate the experiment about multi-objective portfolio problem, thus obtains the optimal investment combination decisions to meet the investor’s requirement, and to some extent, the result improves utility of investment decision, and makes investors to choose more suitable investment decisions for themselves according to personal preference, meanwhile it verifies efficiency and practicality of the improved algorithm respecting to the basic AFSA.In this paper, the strategy of using AFSA to settle multi-objective portfolio problem, is a meaningful attempt in the financial sector, and further broadens application areas of the algorithm.