IA-AIS: An improved adaptive artificial immune system applied to complex optimization problems

This paper proposed an improved adaptive artificial immune system (IA-AIS) for complex optimization problems in continuous search space. In this IA-AIS optimization, several operators are improved or revised which aim at faster convergence speed and better optimal solution. Further speaking, cloning and reproduction of each offspring candidate antibody are proportional to the power of its parent affinity from the antigen; while mutation of each offspring candidate antibody is inversely exponentially determined by its parent affinity from the antigen. Also, suppression operator between antibodies is dynamically controlled according to their concentration. In other words, the suppression level is proportional to their Euclidian distance in continuous search space. The effectiveness of these improvements of operators is experimentally verified. Furthermore, comparative investigations are carried out between the proposed IA-AIS optimization and other optimization utilities. Finally a persuasive case about the proportional-integral-differential (PID) controller tuning demonstrates the potential searching capability and practical value of IA-AIS optimization.

► Clone of antibody is proportional to the power of its parent affinity
► Mutation of antibody is inversely exponentially determined by its parent affinity
► Suppression between antibodies is dynamically controlled by their concentration
► The effectiveness of IA-AIS is experimentally verified on two benchmarks
► IA-AIS performs well in tuning PID controller parameters