An immune memory clonal algorithm for numerical and combinatorial optimization
Ruochen Liu,Licheng Jiao,Yangyang Li,Jing Liu
DOI: https://doi.org/10.1007/s11704-010-0573-6
2010-01-01
Frontiers of Computer Science in China
Abstract:Inspired by the clonal selection theory together with the immune network model, we present a new artificial immune algorithm named the immune memory clonal algorithm (IMCA). The clonal operator, inspired by the immune system, is discussed first. The IMCA includes two versions based on different immune memory mechanisms; they are the adaptive immune memory clonal algorithm (AIMCA) and the immune memory clonal strategy (IMCS). In the AIMCA, the mutation rate and memory unit size of each antibody is adjusted dynamically. The IMCS realizes the evolution of both the antibody population and the memory unit at the same time. By using the clonal selection operator, global searching is effectively combined with local searching. According to the antibody-antibody (Ab-Ab) affinity and the antibody-antigen (Ab-Ag) affinity, The IMCA can adaptively allocate the scale of the memory units and the antibody population. In the experiments, 18 multimodal functions ranging in dimensionality from two, to one thousand and combinatorial optimization problems such as the traveling salesman and knapsack problems (KPs) are used to validate the performance of the IMCA. The computational cost per iteration is presented. Experimental results show that the IMCA has a high convergence speed and a strong ability in enhancing the diversity of the population and avoiding premature convergence to some degree. Theoretical roof is provided that the IMCA is convergent with probability 1.