Research On Portfolio Selection Model Considering Background Asset Risk And Corporate Social Responsibility And Algorithm

In recent years, due to the diversification of financial market risk and the complexity of investment activities, the risk tolerance and financial balance of investors is becoming growing attention. Moreover, with the rapid development of social economy, the contradiction between social, environmental and ethical is growing. After the Markowitz’s mean-variance portfolio selection model, many scholars developed this model according to his thinking. At present, many scholars focused on the financial return and financial variance. And yet, other factors of portfolio selection model are less considered such as the background asset risk of investors and corporation society responsibility. This paper study portfolio problems with background risk and/or corporation society responsibility using probability theory, fuzzy mathematics, the optimum theory and artificial intelligence algorithms, and construct the corresponding theory frameworks and mathematics models. The main researches and contributions of this thesis can be summarized as follows:First, we propose a random portfolio model considering background risk and fuzzy portfolio model considering background risk respectively, and design an improved chaotic fruit fly optimization algorithm. The traditional portfolio models do not consider the background risk which is unable or difficult to trade in financial markets. And more, the objects of the research are often the stock of the return and the variance of portfolio and the related problems, but the increment of these items and the related issues is less researched.Therefore, we establish a random portfolio model considering background risk and fuzzy portfolio model considering background risk which involving the stock and the increment of return, variance, and liquidity, then define corresponding three s-shaped logistic membership functions, use the intersection of the three functions to measure the investor’s subjective satisfaction, and research how the stock and the increment of return, variance, and liquidity impact on investors’ satisfaction. Finally, in order to solve the two portfolio models, on the basis of the original fruit fly optimization algorithm, we initialize the fruit fly population using chaotic operator, and an improved chaotic fly optimization algorithm is proposed.Second, we propose an expected value multi-level portfolio model considering social responsibility and a dependent-chance programming multi-level portfolio model, and design a hybrid fruit fly optimization algorithm based on cloud model. In the process of actual investment decisions, in general, is a layered decision-making process: decision-makers include a leader and a number of subordinates. We assume that the leader and their subordinates have their own decision variables and objective functions. The leader caninfluence on subordinates through his or her behavior. In order to achieve their targets, the subordinates have to decide how to make decisions. And then, these decisions will affect their leader and any other subordinates. Therefore, we propose an expected value multi-level portfolio model considering social responsibility and a dependent-chance multi-level portfolio model. Then, we solve the Nash equilibrium between leader and subordinates using genetic algorithm and neural network.Third, we propose a dynamic portfolio model considering corporate social responsibility and risk aversion coefficient and a portfolio model considering corporate social responsibility and cardinality constraint, design a hybrid fruit fly optimization algorithm based on cloud model. In the actual investment activity, investors need to adjust the portfolio positions in each period when the market environment makes a change in order to maximize the total revenue or minimize total risk at the end of the last period of investment. And then, in actual financial market, due to the costs such as stamp duty and fees derived from the process of security trading, investors can’t hold a lot of the securities at the same time. Therefore, we propose a dynamic portfolio model considering corporate social responsibility and risk aversion coefficient and a portfolio model considering corporate social responsibility and cardinality constraint. Finally, in order to solve the above two portfolio models, we use the cloud operator, crossover operator and mutation operator and design a hybrid fruit fly optimization algorithm based on cloud model.Forth, we propose a random multi-objective portfolio model considering corporate social responsibility and background risk and a fuzzy multi-objective portfolio model considering corporate social responsibility and background risk. As the diversity of sources of investment risk and social, environmental, ethical problems are becoming complicated, investors need to consider their background risk and corporate social responsibility. And more, not only is there variance risk in financial markets, there are risks derived from skewness and kurtosis.Therefore, we propose a random multi-objective portfolio model considering corporate social responsibility and background risk and a fuzzy multi-objective portfolio model considering corporate social responsibility and background risk. In the case of the securities and background risk follow random distribution and normal fuzzy distribution, we deduce the skewness formula and the kurtosis formula of portfolio with background of risk, respectively.In order to solve the above portfolio models, we use the non-dominant sorting operator, the crossover operator and mutation operator and design a hybrid fruit fly multi-objective optimization algorithm based on non-dominant sorting.