This chapter presents the application of the human immune system model and approach to the design of self-organising architectures for artificial systems. The human immune system is a decentralised intelligent system exhibiting remarkable capabilities for learning, information processing and adaptation to environmental changes. Furthermore, it provides an excellent model of adaptive operation at local level and emergent behaviour at global level. There exist several theories aiming to explain immunological phenomena occurring in the human immune system, and numerous software models suitable for their simulation. In this chapter, the idiotypic network model of Jerne is analysed aiming to identify the required elements for implementing it in an agent architecture. The resulting mechanism is illustrated in the collaborative decision-making process carried out by a multi-agent system used for robot soccer simulation.
KeywordsAdaptive Immune System Artificial Immune System Human Immune System Agent Architecture Robot Soccer
- 1.Bersini, H., Varela, F.J.: Hints for adaptive problem solving gleaned from immune networks. In: PPSN, pp. 343–354 (1990) Google Scholar
- 3.Dasgupta, D., Forrest, S.: Tool breakage detection in milling operations using a negative-selection algorithm. Tech. Rep. CS95-5 (1995) Google Scholar
- 8.Jerne, N.: Towards a network theory of the immune system. Ann. Inst. Pasteur., Immunol. 125C, 373–389 (1974) Google Scholar
- 9.Kim, Y.H.: Micro-robot world cup soccer tournament. KAIST (1996) Google Scholar
- 10.Suzuki, J., Yamamoto, Y.: A decentralized policy coordination facility in openwebserver. In: Proceedings of SPA2000 (2000) Google Scholar
- 11.Watanabe, Y., Ishiguro, A., Uchkawa, Y.: Decentralized behavior arbitration mechanism for autonomous mobile robot using immune system. In: Books Artificial Immune Systems and Their Applications, pp. 186–208. Springer, Berlin (1999). ISBN 3-540-64390-7 Google Scholar