A Methodology for Design of Decision Support Systems Based on the Human Information Processing System.

Abstract

This dissertation presents a methodology for the design of Decision Support Systems (DSS). Such a methodology is needed because there is extensive dem and for decision support systems, and existing computer system design methods are not appropriate for these systems. The methodology developd here facilitates decomposing a design problem based on support for the user's information processing system rather than based on problem data or process as in traditional design approaches. It is suitable for use with semi-structured problems, and is usable by systems analysts without graduate levels training. It produces systems which can be implemented quickly using a prototyping approach. The methodology consists of a framework for classification of support alternatives, a step-by-step guide to selection of alternatives appropriate to a user and problem combination, and a breadboard approach to prototyping the alternatives so selected. An empirical study showed that working programmer analysts with no formal training in management science and no previous training in DSS design could use the methodology to produce implementable results. A user's guide to the methodology is presented in an appendix.