A grid-facilitated AIS-based network scheme for many-objective optimization

Abstract

Artificial Immune Systems (AIS), one of the promising artificial intelligence methods, has been widely adopted in the optimization domain. However, their application to many-objective domain is rather scattered. In this respect, we extend the AIS-based algorithm to many-objective situations using the immune network theory facilitated by the grid technique. The network operations are employed not only for managing the diversity, but also to strengthen the exploitation and exploration pressure. The suppression-triggered activation and the archive-driven activation are both introduced in this study to exploit the promising region and to explore along the local Pareto-front. In addition, Grid technique is introduced to reduce the computation complexity in the identification process of the sensory range. Coupled with the grid-facilitated network scheme, the proposed algorithm improves the exploitation and exploration capability in many-objective optimization problem. © 2015 ACM, Inc.