This paper proposes a new complex artificial immune system based on biological immune system. The complex data representation and complex expressions of weights are introduced into the traditional artificial immune system, in which binary or real value data representation was used. The proposed complex artificial immune system imitates the immune response mechanism closely and incorporates the complex partial autocorrelation coefficients of input antigen as the antigen presentation. Simulations of pattern recognition illustrate that the proposed complex artificial immune system has not only the complex memory capacity and classification ability, but also the ability of recognizing translation patterns in high accuracy, which binary artificial immune systems do not have.

complex, artificial immune system, immune response, pattern recognition, translation recognition