Developing integrated decision support systems from mathematical models

摘要

A methodology for the creation of decision support systems (DSS) from mathematical programming models is examined. This approach is demonstrated using a model for flight scheduling, integrating a formal data model, represented in a database management system (Paradox), and a mathematical programming model, represented in an executable mathematical modeling language (GAMS). The integration is completed by creating a database capable of supporting the GAMS model creating a series of queries to extract the data required by the GAMS model, and finally by modifying the GAMS model to import the external data sets and write the results to a file that can be interpreted by the database. This process is first completed manually to support a specific GAMS model. In the second portion of this thesis, the process is generalized to provide a framework that can be used to design a front-end database for any GAMS model. Benefits of this integrated approach over using a stand alone mathematical model include: The assurance of model integrity, explicit data modeling, improved representation and manipulation of model inputs and outputs, greater integrity of input data, and easier interpretation and multiple views of model outputs.