An artificial immune network with diversity and its applications

Living body immune system is highly complicated and appears to be precisely tuned to the problem of detecting and eliminating infections. Diversity is an important source of robustness in the living body immune system. Based on immunological principles, we propose an artificial immune network with diversity inspired by the living body immune system and apply it to pattern recognition. We express the diversity of living body immune system by the fact that minute random noise is added to the proposed artificial immune network repeatedly and test the proposed artificial immune network by the simulation on alphabet pattern recognition. The simulation results illustrate that the proposed artificial immune network with diversity is effective in learning a stable recognition code in response to an arbitrary sequence of continuous valued input patterns as well as binary input patterns and it is able to improve the pattern recognition capability and noise immunity in the noise environment.