OBJECT-ORIENTED MODEL INTEGRATION IN MIDAS

Decision support systems for users without modelling expertise require domain-specific modelling knowledge to translate between conceptual and mathematical problem views. In complez, dynamic decision situations using multiple model types, the system must create and modify individual models to reflect changing domain conditions and assumptions, maintain consistency among different models in the same decision situation and allow communication among models for their complementary use, all without relying on user expertise. This paper describes a system for debt decision support which flexibly integrates optimization and simulation modelling and heuristic reasoning for non-expert users through an object-oriented, domain-specific knowledge base. Stable domain relationships and mathematical procedures are encapsulated in domain object classes; domain object instances are combined to form common model representations manipulated by operators specific to each model type. The approach is applicable in domains in which stable entities and interactions exist and in which model flexibility results from varying combinations of entities-conditions which are found in many financial and other business modelling situations.

[1]  Patrick Henry Winston,et al.  Amplifying Expertise with Expert Systems , 1986 .

[2]  Matthias Jarke,et al.  Logic-Based Formula Management Strategies in an Actuarial Consulting System , 1984, Decis. Support Syst..

[3]  Basil A. Kalymon.  Bond Refunding with Stochastic Interest Rates , 1971 .

[4]  Robert E. Filman,et al.  Reasoning with worlds and truth maintenance in a knowledge-based programming environment , 1988, CACM.

[5]  Vasant Dhar,et al.  Rule-based versus structure-based models for explaining and generating expert behavior , 1987, CACM.

[6]  Vasant Dhar,et al.  Knowledge-based decision support in business: Issues and a solution , 1988, IEEE Expert.

[7]  W. Boyce,et al.  Optimum Bond Calling and Refunding , 1979 .

[8]  Daniel G. Bobrow,et al.  Object-Oriented Programming: Themes and Variations , 1989, AI Mag..

[9]  Jeremy F. Shapiro,et al.  Stochastic programming models for dedicated portfolio selection , 1988 .

[10]  Jay F. Nunamaker,et al.  Model Management Systems: Proposed Model Representations and Future Designs , 1985, ICIS.

[11]  Marvin Minsky,et al.  A framework for representing knowledge , 1974 .

[12]  Matthias Jarke,et al.  Domain-Specific Dss Tools for Knowledge-Based Model Building , 1985, Decis. Support Syst..

[13]  Richard Fikes,et al.  The role of frame-based representation in reasoning , 1985, CACM.

[14]  S. Bradley,et al.  A Dynamic Model for Bond Portfolio Management , 1972 .

[15]  Robert H. Bonczek,et al.  A Generalized Decision Support System Using Predicate Calculus and Network Data Base Management , 1981, Oper. Res..

[16]  Joyce J. Elam,et al.  USING ARTIFICIAL INTELLIGENCE TECHNIQUES TO ENHANCE THE CAPABILITIES OF MODEL MANAGEMENT SYSTEMS , 1987 .

[17]  Arthur M. Geoffrion,et al.  An Introduction to Structured Modeling , 1987 .

[18]  Robert W. Blanning.  A relational theory of model management , 1987 .