Interactive multi-criterion decision-making method and device including preference aid decision factor optimization and computer equipment

Abstract

The invention discloses an interactive multi-criterion decision-making method containing preference auxiliary decision-making factor optimization, and the method comprises the following steps: settingthe number of interactive iterations based on the operation of an evolutionary multi-objective optimization algorithm, and obtaining a population initial solution of investment portfolio double problems; making an interactive preference feedback mode algorithm based on a preference auxiliary decision factor interact with a decision maker to obtain decision maker preference feedback; establishinga preference model based on the decision tree of the classification rule, and obtaining model preference feedback by utilizing the preference model; calculating an evaluation preference model by utilizing decision maker preference feedback and model preference feedback; and training and updating the preference model until a training termination condition is satisfied. The invention further discloses an interactive multi-criterion decision-making device containing preference auxiliary decision-making factor optimization, and the interactive multi-criterion decision-making device can obtain higher average prediction accuracy and can obtain a satisfactory solution with more centralized distribution.

Description

technical field [0001] The invention relates to the technical field of target optimization in the field of investment portfolio, in particular to an interactive multi-criteria decision-making method, device and computer equipment including preference auxiliary decision-making factors optimization. Background technique [0002] In the investment portfolio optimization problem, the decision-maker is most concerned about how to achieve the optimal allocation of limited funds among a certain number of assets (such as stocks, bonds), so as to obtain as high a return as possible and as much as possible at the end of the investment period. Possibly low risk. The Mean-Variance (MV) portfolio formula first proposed by Nobel Prize winner Harry Markowitz has made great contributions to the decision-making of capital investment in modern computational finance. The portfolio formula has two optimization objectives: to maximize the return of the portfolio (calculated by the mean of the e...

AI Technical Summary

Problems solved by technology

[0005] According to the theory of bounded rationality, decision makers often do not have a clear understanding of optimization problems, and in current related research, decision makers can only make preference judgments based on the value of the objective function of the optimization problem, which means that decision makers may lack sufficient knowledge when making judgments. Associated information

Therefore, although the difficulty of pairwise comparison is relatively low, decision makers may give inconsistent judgments, uncertain judgments (that is, they cannot judge whether a pair of solutions are good or bad) or conflicting judgments due to uncertainty in the decision-making process. Judgment (such as A is better than B, B is better than C, C is better than A), that is, feedback noise is generated, and the resulting preference information cannot truly reflect its preference tendency

Although some scholars have pointed out the importance and research prospects of controlling feedback noise, they have not yet proposed an effective solution

Images