<< Click to Display Table of Contents >> Navigation:  Decision-theoretic modeling > Changes in structure

Changes in structure are external interventions that modify a system in question. An example of a change in structure is imposition of a new tax within the economic system of a country. It is of critical interest to decision makers to be able to make predictions of the effects of changes in structure. Since the model reflects the reality in its unmanipulated form and changes in structure of the kind contemplated by the decision maker have perhaps never been performed, predicting their effect is in general daunting.

In order to be able to predict the effect of arbitrary changes in structure, it is necessary that the model contain causal information. Directed graphs allow for representation of causality. One may adopt the convention that each arc in the graph denotes a direct causal relation between the parent and the child node. We recommend that all models built are causal in that sense. The operation of controlling a value is an example of a causal manipulation and may result in changes in structure.

GeNIe and SMILE are unique in supporting changes in structure in decision models. Please see Controlling values of variables section of this document for additional details.