This article examines the application of a modeling technique called influence diagramming to decision support system (DSS) analysis. Influence diagrams allow the development of a model of the natural (at least from the perspective of the decision maker) structure of a decision situation. These diagrams are simple to develop and understand, providing the DSS analyst and the decision maker with a useful basis for communication. The diagrams can also be easily translated into specifications detailing the design of a computer-based system. In particular, influence diagrams are well suited to the specification and description of spreadsheet-based models commonly used in contemporary decision support systems.
Key Terms in this Chapter
Influence Diagram: A visual representation of a problem showing the variables involved and the influence relationships between the variables. They can be used in decision analysis to calculate expected values or in descriptive modeling providing a conceptual tool for thinking about a problem prior to model development using a computer-based tool like a spreadsheet.
Decision Node: A variable included on an influence diagram or decision tree that is under the control of the decision maker. It is usually represented visually as a square or rectangle.
Chance Node: A variable included on an influence diagram the decision maker does not control. The possible values of the node are usually associated with some probability distribution. Chance nodes are usually environmental variables.
Bubble Diagram: An alternative name for an influence diagram used by the authors of a well-known guide to spreadsheet development (Read & Baston, 1999).
Outcome Node: A variable included on an influence diagram that represents a measure that is included in the problem objective. For example, in a financial problem with a goal to increase profit, the outcome node might be a variable called profit.
Intermediate Node: A variable included on an influence diagram to show intermediate results.
Decision Analysis: A structured approach to problem solving that attempts to apply mathematical models to help structure and solve problems.
Exogenous Node: A variable included on an influence diagram that has an impact on the problem being studied but is considered to be outside the system. It is a special form or a chance node. It is usually represented visually as a double lined circle.