The availability of microcomputers, modeling langauges and general purpose spreadsheets has resulted in an increase in the use of models for decision making within organizatons. Decision makers with microcomputers on their desks and spreadsheet and modeling software can create models rapidly. Problems with model redundancy, consistency, integrity and security have prompted an increased interest in the design of model management systems (MMS). Several model management designs have been discussed in the literature. Different model representation techniques have been proposed. These include formal logic, semantic inheritance networks, frames, and relational representations. The approaches to model management are evaluated in respect to their model manipulation and model storage functions. A framework for the design of MMS is proposed based on the system design objectives and the system domain complexity. Advantages and disadvantages of each model representation method are identified. Application domains for the classifications are proposed which focus on the strengths and weaknesses of the model representation for supporting model storage and model manipulation functions. An example of the design of a MMS using the classification is presented. Introduction Model management systems (MMS) have been proposed as a possible solution to the need for support of the The use of models within organizations has increased modeling activities of organizations (Will, 1975). An dramatically over the past few years (Ulvila and Brown, MMS is a software system which provides for the cre1984). The availability of microcomputers, modeling ation, manipulation and access of models. The ability to languages (e.g., IFPS), and general purpose spreadstore and manipulate models is a critical function in the sheets (e.g., Supercalc3, Lotus 1-2-3) has provided decidesign and implementation of an MMS. Model storage sion makers with the necessary development tools for functions include model representation, model abstracrapid model implementation and design. These models, tion, physical m odel storage and logical model storage. developed by different decision makers, can lead to difModel manipula tion functions include model instantiaferent and often conflicting results. Problems of model tion, model selection and model synthesis. The MMS redundancy, inconsistency, integrity and security have accesses the centralized database of the organiz ation and been noted. These problems are similar to the data manexternal databases to obtain the data necessary to solve a agement problems that prompted the design of database given problem. The capability for interactive input and management systems. correction is also provided. Two objectives for model