DESIGN OF A MANAGEMENT INFORMATION SYSTEM
Abstract
Current data base design has been a modeling approach in which an experienced designer models the environment by identifying entities and their relationships. A data base schema is then created according to what data base model is used. Two drawbacks in this approach are (1) data base performance relies on the designer's ability to perceive the environment and (2) data base design has not been automated. In this study, we present a data base design methodology based on a set of user specified reports. The goal is to design a data base which is capable of supporting or generating these reports. Therefore, it is a functional approach rather than a modeling approach. An Artificial Intelligence's production system is intoduced to simulate human intellectual activity generally and data base design specifically. It is an automated procedure and no modeling ability is required. Furthermore, the production rules in a production system can be easily modified to adapt to any changes in environment requirements or data base technology such as data base restructuring and interactive data base design. To emphasize that data processing is as important as data storage and data retrieval in information processing activities, we also introduce procedures in a data base for users' convenience. The goal is to have a system which retrieves data and executes procedures automatically without requiring a user to write programs. We discuss the theoretical issues concerning automatic data retrieval and data processing in the framework of the extended first order logic and the resolution principle. In particular, the relationship between Skolem functions and procedures is exploited and utilized in resolution for automatic data processing. We also suggest data base partitioning as a means to improve efficiency which is usually a problem encountered in most of the resolution based systems. Data base partitioning, defined by the second order logic expressions, refers to the division of a data base into subsets such that a query only has to be processed in a small subset rather than the entire data base. Finally, we discuss the organization and the structure of procedures for the extended data base.
Degree
Ph.D.
Subject Area
Management
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.