A Hybrid Artificial Immune Network with Swarm Learning

The artificial immune system is a new approach inspired from abundant mechanisms of biological immune system. It includes such basic operations as clone, mutation, and selection, even crossover. It is widely applied to function optimization, abnormal detection, pattern recognition, computer security, machine learning, control engineering, etc. However, the evolutionary process of the current artificial immune system depends on only two factors. One is the fitness between antibody and antigen, and the other is the concentration of antibody population. As a global searching method, particle swarm optimization includes an important social learning mechanism that enables it to fast approximate the global optimum. This paper proposed a hybrid artificial immune network for optimization with swarm learning and elite-keeping. Simulation results indicated this hybrid method has lower time complexity and fast convergence, and is an effective optimization tool.