Dempster-Shafer theoretic design of a decision support system for a large-complex system

This paper discusses the design of a decision support system (DSS) which can process and display uncertainty contained in information regarding operating conditions of the plant. The authors take a probabilistic approach to evaluate effectiveness of a decision support system. Either of the plant state information provided by the DSS and operator's opinion regarding the plant state is represented in terms of a belief structure. These two belief structures are combined into one by the operator. Criteria for determining a control action based on the combined belief structure are given. The authors classify the operator into three types according to his decision criterion. The authors analyze conditions for which the DSS can provide the operator with a support so that the expected loss of the plant may be reduced when a combination rule is given and the operator's knowledge of characteristics of the plant is assumed.<<ETX>>