Construction of a model formulation consultant: the AEROBA experience

Abstract

Typically, most of the current decision support systems (DSSs) do not possess a module to help users formulate mathematical programming models. This observation has led researchers to offer different approaches to the design of formulation consultants. Generally, these approaches emphasize the automation of formulation tasks with a predetermined (static) control or guidance strategy. A new approach to constructing a formulation consultant that emphasizes dynamic formulation control, opportunistic control, is offered. A system, AEROBA, has been developed using a blackboard architecture to provide support to the formulation process. Several of the issues raised regarding model formulation, opportunistic control, and blackboard systems are addressed using the AEROBA implementation. The contributions of this research are: (1) understanding the process of model formulation; (2) addressing the concerns of model formulation in attempting to automate the process; and (3) using the blackboard paradigm to deal with the unique characteristics of the model formulation process