The AIS-SoL Optimization: An Artificial Immune System with Social Learning

This paper proposes an artificial immune system with social learning (AIS-SoL) for optimization. In the AIS-SoL optimization, the candidate antibodies is separated into two levels i.e., the top level of elitist antibodies (TLEA) and the bottom level of common antibodies (BLCA). Different level of antibodies experience different mutation processes, i.e., a self-learning strategy is used for the TLEA while a social-learning strategy is applied to the BLCA. According to the social-learning strategy, each antibody in BLCA learns from an elitist antibody randomly selected from the TLEA. Some numerical simulations are arranged to evaluate the performance of the proposed AIS-SoL. The results demonstrate that the proposed AIS-SoL optimization outperforms the canonical opt-aiNet, the IA-AIS and the AAIS-2S in both convergence speed and solution accuracy.