An Approach to Combine Artificial Immune Systems and Neural Networks: A Dynamic Artificial Immune System Based on the Self-Organizing Map

Computational Intelligence aims at creating problem-solving strategies which are inspired by biological systems. From this approach emerged artificial neural networks, genetic algorithms and recently artificial immune systems. Comparing artificial immune systems and genetic algorithms, close computational similarities can be found. Furthermore, research showed the biological nervous and endocrine systems to be so intertwined that science now speaks of one single system, the neuroendocrine system. From this perspective, the question emerges whether the interrelations between the nervous, endocrine and immune system can yield new or more general computational systems that can be used for problem-solving as well. In the present work, an approach to combine artificial neural networks and artificial immune systems is introduced together with a biological foundation of the computational processes shared by both paradigms. On that basis, an artificial immune system that rests upon the self-organizing map is implemented and demonstrated by simulations.