Stochastic multi-attribute decision-making based on double-reference points for selecting informatization alternatives in agricultural enterprises

摘要

Existing studies on stochastic multi-attribute decision making (SMADM) are mainly conducted from a single reference point. This approach fails to consider the duality of reference points. The present study solves the SMADM problem with double reference points based on bounded rationality hypothesis. The characteristics of double reference points and their influence on decision behavior are described using behavioral decision theory. This study refers to behavioral psychology and the science of decision making to illustrate the relationship between double reference points and loss. Decision making in informatization projects for agricultural enterprises is used as example. Three unsatisfactory degree functions are established and the loss function is introduced. This approach facilitates the ranking and selection of alternatives by aggregating the decision information with expected loss according to the principle of loss aversion. The scientific nature of the new method is explained using informatization projects in agricultural enterprises as example. Results show that the relationship between attribute value and reference points directly affects the decision maker, whose decision behavior is mainly based on loss value. Such a value is primarily influenced by the attitude of the decision maker. A highly positive correlation exists between unsatisfactory degree and loss value, whereas varied "unsatisfactory degrees" directly determine the types of loss. When "unsatisfactory degrees" are the numbers within (0,1), the greater the unsatisfactory degrees, the greater the loss. The magnitude of loss is directly affected by the type of loss and its conversion rate. These results can help solve decision making problems in agricultural enterprises and enrich the research categories for SMADM problems. This development provides reference for the research of multi-attribute decision making with multiple reference points.