The Model Management

The model management component of a DSS, consists of analytical tools, such as statistical models, financial models, artificial intelligence heuristics, and operations research models, as well as a function for managing those tools. Some of the tools consist of prepackaged analyses, while others provide the users the opportunity to build their own models. The value of this component results from providing easily implemented access to a wide variety of tools and assistance in using the tools, so the users can and will investigate relevant patterns and issues in their data. Hence, the goal is to enable users to select the models they perceive as most appropriate to assist with the particular question under consideration. This goal is only achieved, however, if DSS are designed appropriately for the individuals or groups who will be using them.

Clearly this is not a problem if all questions and all data have a unique modeling opportunity associated with them. That is, if one believes there is only one way to analyze data correctly, then the transnational nature of a DSS should not affect the design of the model management system. However, that assumption is rarely correct. Even if one simply acknowledges that different divisions will have different perspectives that will affect their approach to decision making, it becomes obvious that they will need to consider different data in different ways to address those differences in perspectives. Hence, the various divisions will require different (and perhaps vastly different) models to support those decisions. In addition, since management style is at least partially a function of the state of development and technology, the variations in these factors will increase the heterogeneity of models required of a successful system.