The Application of Recognition Algorithm Based on Artificial Immune and the Study of Its Convergence

A new artificial immune recognition algorithm based on biological immune system is proposed in this paper and its convergence is proved. The algorithm has the following features: calculating the affinity between antibodies and antigens with improved Euclid distance; increasing the variety of antibodies by adopting antibody mutation, which also enhances the ability of recognizing antigens; finally antibodies memory matrixes come into being, which results in the quick recognition on antigens. Fixed point theory and compressed mapping theorem are used to prove the convergence of the algorithm. It analyzed the application of this algorithm in state recognition of a heating furnace, it proved that the algorithm has high recognition rate and high learning rate