DOMAIN-SPECIFIC DSS TOOLS FOR KNOWLEDGE-BASED MODEL BUILDING

Abstract

The formulation of complex planning models, such as linear programming (LP) systems, is a difficult task that enjoys little support by current decision support systems tools. It is hypothesized that current artificial intelligence technology is insufficient to build generalized formulation tools that would be usable by OR-naive end users. As an alternative, this paper presents a domain-specific approach to knowledge-based model formulation which combines the use of "syntactic" knowledge about linear programming with “semantic” guidance by knowledge specific to some application domain. As a prototype of this approach, a model formulation tool for LP-based production management is under development at New York University.