A formal analysis of the model management literature

This paper presents an extensive bibliography of the archival material on model management, and an analysis of that material. It argues that the three general approaches to model management can be observed, called here the data-model analogy, knowledge-based and model-representation approaches. For the period 1970 to 1989, 57 journals were exhaustively searched for articles dealing with model management; 152 articles were identified. An informal analysis shows that model creation and execution have been the major model management functions addressed by research, and that mathematical programming and simulation have been the dominant model areas. Applications have addressed the extremes of Operations Research models, namely policy and strategy at one extreme and operations at the other, with many applications in accounting and finance also prevalent. A formal analysis using factor analysis shows that research in model management has moved away from speculation and focus on certain limited aspects of the subject towards the practical implementation of model management systems. Further, a regression model shows that the model-representation approach has tended to yield implemented systems, whereas the data-model analogy approach has not. Directions for future research and development are suggested.