Back to the program.
An Engineering Approach to Shape-Space - Making the Most of Artificial Immune Systems
Emma Hart
Napier University
Scotland
e.hart@napier.ac.uk

Abstract

The field of Artificial Immune Systems has demonstrated significant advances over the last decade resulting in systems which have been applied to a diverse range of scenarios, including for example intrusion detection, optimisation, data-mining and condition monitoring. These diverse application areas have been tackled using an equally diverse range of immune-inspired techniques. However, appraisal of these successes should be tempered somewhat by the facts that: 1) The algorithms are based on very naive extraction of immune metaphors 2) there is generally little theoretical justification to support any of the proposed algorithms, and 3) there is a lack of understanding of which techniques are appropriate in which situations. This talk describes an attempt to redress the balance by taking a "back-to-basics" approach. We implement a visual simulator to investigate the role that is played by shape-space, a key ingredient of any immune metaphor. By providing a simple way to manipulate shape-space and its parameters and understand the effects of such manipulations, we begin to build an understanding of how as engineers, we can really exploit immune-algorithms by specifically tailoring them to our own applications.

Keywords

Artificial Immune Systems