Structure frame-based model management system
Jae Sik Lee, Monique Guignard, Christopher Jones · Medical Entomology and Zoology · 1989
A relatively unexplored area of model management systems research is the automatic construction of algorithms to solve a given model. In order to construct or suggest an appropriate algorithm, it is necessary to analyze and identify the structure of the given model. In this dissertation, we present an automatic procedure for identifying the structures of mathematical programming models expressed in algebraic form. We have developed techniques for extracting structural (syntactic and semantic) information from the algebraic form, techniques for storing this structural information, and finally, techniques for analyzing this information to identify the problem that the model represents. Storing and identifying a model by its structural information makes it possible to automatically suggest an appropriate algorithm for solving the model. We demonstrate the practicality of our approach with a prototype implementation called the Structure Frame based Model Management System.