A solution concept for artificial immune networks: A coevolutionary perspective

Abstract

In this paper a relation between artificial immune network algorithms and coevolutionary algorithms is established. Such relation shows that these kind of algorithms present several similarities, but also remarks features which are unique from artificial immune networks. The main contribution of this paper is to use such relation to apply a formalism from coevolutionary algorithms called solution concept to artificial immune networks. Preliminary experiments performed using the aiNet algorithm over three datasets showed that the proposed solution concept is useful to monitor algorithm progress and to devise stopping criteria. © Springer-Verlag Berlin Heidelberg 2007.