Decision model

From Wikipedia, the free encyclopedia

A decision method is a formal (axiomatic) system that contains at least one action axiom. An action is of the form "IF <this> is true, THEN do <that>". An action axiom tests a condition (antecedent) and, if the condition has been met, then (consequent) it suggests (mandates) an action: from knowledge to action. Excepting very simple situations, successful action axioms are used in an iterative manner. For example, for decision analysis, the sole action axiom occurs it the Evaluation stage of a four-step cycle: Formulate, Evaluate, Interpret/Appraise, Refine.

Formulation

Formulation is the first and often most challenging stage in using formal decision methods (and in decision analysis in particular). The objective of the formulation stage is to develop a formal model of the given decision.

Evaluation

Evaluation is the second and most algorithmic stage in using formal decision methods. The objective of the evaluation stage is to produce a formal recommendation (and its associated sensitivities) from a formal model of the decision situation.

Appraisal

Appraisal is the third and most insightful stage in using formal decision methods. The objective of the appraisal stage is for the decision maker to develop insight into the decision and determine a clear course of action. Much of the insight developed in this stage results from exploring the implications of the formal decision model developed during the formulation stage (i.e., from mining the model). Central to these implications is the formal recommendation for action calculated during the evaluation stage. Other implications include various forms of Sensitivity of the recommendation to various components of the decision model. Insight may also result from discussion of the key aspects of the reasoning that led to the formal decision model (i.e., by justifying the model). Possible actions following the appraisal stage include implementing the recommended course of action, revising the formal model and reevaluating it, or abandoning the analysis and doing something else.

Justifying a decision model entails exploring and explaining the reasoning that led to the formulation of particular aspects of the decision model.

Mining a decision model entails extracting information (e.g., sensitivity, value of prediction, and value of revelation) from a given decision model.

Refinement

Refinement is the fourth and most critical stage in using formal decision methods. The refinement stage responds to the insights obtained during the Appraisal stage. Effective refinement activities include opportunities to test possible decision model changes to see their implications and suggest better ways to modify the decision model.

References

Holtzman, S. 1989. Intelligent Decision Systems. Addison-Wesley.

This article is issued from Wikipedia. The text is available under the Creative Commons Attribution/Share Alike; additional terms may apply for the media files.