Making intelligent agents more intelligent

Lin Padgham

School of CS&IT, RMIT University

Date and time: 11.30am-12.30pm, Friday 3rd September, 2004

Venue: 10.11.04

Chair: Michael Winikoff

Abstract:

This seminar describes work within the intelligent agent group that looks at enhancements that can be made to the popular BDI (belief, Desire, Intention) agent infrastructure, to support building smarter agents. BDI systems already support a number of features which facilitate building of powerful agent systems. However there are (of course) also significant limitations. While arbitrary amounts of additional intelligence can be built into an agent system on an application by application basis, we are interested in generic mechanisms that can provide well integrated infrastructure support for application developers. This seminar describes work in three different areas that all contribute to this goal of enhancing the infrastructure support for making BDI agent systems more intelligent. The three areas to be covered are: reasoning about concurrent interacting goals, planning, and adaptivity using learning. The work on reasoning about goals will be covered in more detail than the other two areas, as the work is more advanced. The seminar seeks to describe the motivations and benefits that can be realised, and to give an overview of the general approaches developed rather than technical details.

About the speaker:

Lin Padgham is an Associate Professor in the School of Computer Science and Information Technology at RMIT University.

Seminars are free and open to the general public. No booking is necessary.

If you are interested in giving a presentation in this seminar series, or to make suggestions for speakers, please contact James Harland, the seminar co-ordinator.