The Prometheus Methodology

Abstract

In this chapter we present the Prometheus methodology for building agent-based software systems. Our goal in developing Prometheus was to have a process with associated deliverables which could be used by industry practitioners and undergraduate students without a previous background in agents. As a result, the Prometheus methodology aims to be detailed and complete, as well as being general-purpose and having tool support. Prometheus comprises three phases: system specification, architectural design, and detailed design. The Prometheus methodology has been developed over a number of years as a response to both educational and industrial needs. The methodology has been used by industrial practitioners, taught at workshops at a number of conferences, and has been taught to undergraduate and postgraduate students, as well as having been used in student projects. These experiences have been positive and we have noticed an enormous difference in the ability of our students to develop agent systems. Using Prometheus third year undergraduates are able to build reasonable agent systems in a one semester course, something that previously was challenging for graduate students.

Figures 11.1 and 11.2, the Query Late Books Scenario, and some of the text in the Architectural Design section are reproduced by permission of John Wiley & Sons, Ltd. from Lin Padgham and Michael Winikoff, Developing Intelligent Agent Systems: A Practical Guide to Design, ISBN 0-470-86120-7 to be published in 2004.