A unified model for data acquisition and decision making

Abstract

In this paper we present a model for decision making employing both the concepts of expected utility and belief functions. The situation modeled is that of a decision maker trying to select a single course of action from a finite set of possible courses of action. Although this selection is the major decision, the decision maker is also confronted with many preliminary decisions such as which data sources to access to obtain pertinent information and how to evaluate the mforma tion received. The proposed model partitions the total decision making problem into two interlocking sub-processes—the data acquisi tion process and the final decision process which is based on an elimination procedure. Decisions in the data acquisition phase employ a rationale based on expected utility, while the criterion upon which the final decision is made is based on belief functions. The application of this model is illustrated by two related examples which are discussed in some detail. The approach proposed in this paper is in a sense a mathematical structunng of the type of reasoning intuitively followed by an intelligent human decision maker when faced with a single important decision. The mathematical modeling of such processes is becoming increasingly important as the role of the computer in modem decision making continues to grow.