Intelligence Algorithm For Multi-period Uncertain Portfolio Optimization Selection

After Markowitz proposed mean-variance model,the study of portfolio selection problem has attracted wide attention and well investigated.However,most traditional researches are based on the basis of probability theory,and assumed that the security returns are random variables.As is known,the fundamental premise of applying probability theory is that enough historical data should be provided for the estimation of the probability distribution.Since the complexity of the real financial market,we sometimes cannot obtain sufficient historical data.In such situations,we have to rely on the experts' subjective evaluations.Therefore,some researchers study portfolio selection on the basis of uncertain theory and regards return as uncertain variable.Note that most of the researches are mainly focused on single-period portfolio selections.However,in real world,investors often need to dynamically adjust investment strategies according to market.Thus,the purpose of this paper is to study multi-period portfolio selection problems on the basis of uncertain theory.The main research work of the paper is summarized as follows:(1)Assumed the security return as uncertain variable,firstly,a multi-period multi-criteria portfolio optimization model was proposed by considering real-world constraints.For solution,we apply a weighted max-min fuzzy goal programming approach to convert the proposed multi-objective programming model into a single-objective one.After that,we design a novel hybrid of imperialist competitive algorithm and firefly algorithm,termed ICA-FA to solve it.Finally,we provide a numerical example to demonstrate the performance of the presented model and corresponding algorithm.(2)When distributions of security returns are asymmetric,semi-variance was proposed to replace variance as the measure of the risk in portfolio selection.Then,we construct a multi-period uncertain mean-semi-variance portfolio selection model under the real-world constraints.After that,design a novel ICA-GA to solve it.Finally,we provide a numerical example to show the performance of the given model and corresponding algorithm.