Research On Portfolio Models Based On Double Expected Utility With Intelligent Algorithms

The general utility functions do not consider the possibility of future events in the history of the difference,which the probability of occurrence in history will occur with the same probability in the next time,in order to avoid such a strong assumption,this thesis researches portfolio model based on the double the expected utility,and applies artificial intelligence algorithm to solution it.Therefore,the utility function of double expected utility theory to measure the investment behavior,the major work is described as follows:1 According to the needs of the financial markets,the return rate of stock obeys the normal distribution is the premise,lead into the maximum number of investment ceiling,the establishment of multi-objective portfolio model,use penalty function method transformed the multi-objective portfolio model into a single objective model,and construct the model teaching-learning-based algorithm to solve the problem,and select 10 stocks by simulation experiment.2 The introduction of restrictions of short selling,establish a portfolio optimization model,at the same time as the reference weight coefficient of risk aversion factor,which makes the model more in line with the decision of investor psychology,so as to ensure the feasibility of the decision scheme.In addition,this thesis design a chaotic algorithm to solve the model for bird swarm model,and compare to particle swarm teaching-learning-based algorithm,the result of when the risk aversion factor A takes different values,the chaotic bird swarm algorithm in this thesis has better results,and provides a better solution for investment decision makers.3 Considering China's real stock market,there are some friction factors,contain the minimum trading volume and transaction cost,while standing on the investors,we must aim to spread the risks,investment set limit,which makes the model closer to the investors in the financial market are decision-making behavior.This model scheme is reasonable and feasible,so that investors more options;In addition,the solving model of particle swarm algorithm design,the numerical results accord with the theoretical basis,provide an optimal choice for investors.