Based on Differential Evolution Method
Yulian Fan,Yang Xu,Mengying Chang
2018-01-01
Abstract:Portfolio optimization refers to the fact that investors allocate funds to certain kinds of assets, so that the investment amount of each type of assets accounts for a certain proportion of the total investment. The purpose is to make the overall income of the assets held by investors as high as possible, or make the investment risk as low as possible. With the deepening of people's research, the theory of portfolio optimization has been gradually applied into a more successful and more mature theory. This paper intends to use the basic prospect theory model, using Matlab software, based on the differential evolution algorithm, to optimize the real economic market data of 10 assets. Keywords—portfolio optimization; prospect theory model; differential evolution algorithm I. FOREWORD (STUDY BACKGROUND) With the continuous development of China's economy and the deepening of economic system reform, the concept of financial management of Chinese nationals has gradually improved, and investment has gradually become a widely used investment channel. Investing can make us profitable, but investors also have to bear certain risks while gaining income. How can investors reasonably allocate funds to invest in different assets, determine the appropriate proportion of investment in a variety of assets to the total amount of investment, so that the total income of the assets held by investors is as high as possible and the risk is as low as possible. How to calculate the risks and benefits of portfolio investment and how to allocate assets to achieve a certain balance between these two indicators. These are all issues that investors need to solve urgently. In this research context, this paper has conducted in-depth research on portfolio optimization. II. PORTFOLIO OPTIMIZATION PROBLEM A. Introduction The main purpose of a good stock investment is to effectively avoid stock market risks and get the most benefit. Modern asset portfolio theory was first developed by American economist Harry M. Markowit (1927) proposed, he applied the mean and variance of portfolio compensation and found the optimal solution by establishing a mathematical model. The expected utility was first proposed by the Swiss mathematician Bemoulli, and by VonNeumann, Morgenstem et al., based on the answer to the "St. Petersburg Paradox" by Bemoulli, after strict axiomatization assumptions. It describes how “rational people” determine optimal decisions in the face of uncertain situations. The theory of prospects is based on the expected utility, adding important psychological components and considering human behavior in the decision-making process. This paper will use the prospect theory model to calculate the optimal solution of the model through differential evolution algorithm. B. Symbol hypothesis Assuming N is the number of assets and ωi ≥ 0 is the weight of asset i in the portfolio, then a portfolio can be expressed as xi= ω1,⋯,ωN , ∑ ωi i=1 . S is assumed to be the number of situations, ps is the probability of occurrence of s moment, ris is the return of asset i at time s, r̅i is the average return of asset i, and d is the ideal return level. III. BASIC PROSPECT THEORY MODEL The prospect theory is a decisive paradigm decision model proposed by Kahneman and Tversky (1979).It divides risk decisions into two processes: editing and evaluation. In the editing phase, the individual's gains and losses in the portfolio are determined by defining a reference point r0 ,where r0 represents the zero gain or zero loss that the individual believes. In the evaluation phase, the weight function π(p) and the value function υ(r) that depend on the subjective probability are used to judge the information. In the prospect theory, the probability weight function π(p) is to measure the effect of the event on the desirability of the prospect. π(p) is an increasing function, π(0) = 0 , π(1) = 1 , when the probability p is very small, then π(p) ≥ p. In the prospect theory, the value function υ(r) is the behavior value that describes the result of the gain or loss. For a given reference point r0 , the image is usually asymmetrical because the function reflects the attitude of different investors towards gains and losses; for the benefit is upward concave, for the loss is convex downward. Furthermore, typically the value function υ(r) is steeper for the loss than for the gain, for s>0, there is υ( ) ≤− υ( ). In this paper, the value function uses the explicit formula given in Tversky and Kahneman (1992): υ r = (r − r0),if r ≥ r0, − (r0 − r),if r ≤ r0, 3rd International Conference on Communications, Information Management and Network Security (CIMNS 2018) Copyright © 2018, the Authors. Published by Atlantis Press. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/). Advances in Computer Science Research, volume 65