Study And Application On Fuzzy Artificial Immune Network

In nature, living organisms are doted with sophisticated learning and processing abilities that allow them to survive and proliferate generation after generation in their dynamic and competitive environments. For this reason, nature has always served as inspiration for many scientific and technological developments. The information processing is divided for three types, neural system, immune system and excretory system. The natural immune system can be seen as a parallel and distributed adaptive system that has tremendous potential in many intelligent computing applications. This is because the immune system exhibits the following points of strength: recognition, feature extraction, diversity, learning, memory, distributed detection, and self-regulation etc.Inspired by the Immune System, from the angle of pattern recognition, a kind of fuzzy artificial immune network is discussed in this paper. Firstly, introduce some novel algorithms for pattern recognition and point out integrating these algorithms will exploit new area in the development of pattern recognition. Secondly, biologic immune system is introduced and the previous artificial immune network is discussed briefly. Then, refered to existed results (mainly refer to Castro's aiNet), based on the full comprehension of the fuzziness in information processing and the relative mechanism in biologic immune system, a novel algorithm for data clustering and compressing, named Evolutionary Fuzzy Artificial Immune Network (EFAIN), is proposed. In this algorithm, an idiotype fuzzy immune recognition hypersphere as antibody which has certain structure and function in a sense is explored. And the algorithm also simulates the characteristic of secondary immune response in biologic immune system. The related experiments show that this algorithm can be used for clustering and compressing data which has unclear boundary and complex distribution shape. Finally, the problem of how to select the centers of Radial Basis Function (RBF) Network for function approximation and regression is studied. The algorithm is used to this problem after a little change to it. Experiments show that this algorithm is efficient.