Artificial immune system

In 1980s, Farmer and others took the lead in giving a dynamic model of immune system based on immune network theory, and discussed the relationship between immune system and other artificial intelligence methods, and began the research of artificial immune system. It was not until 1996 that an international symposium based on immune system was held in Japan for the first time, and the concept of "artificial immune system" (AIS) was put forward for the first time. Subsequently, the artificial immune system entered a period of prosperity and development. D. Dasgupta and Jiao Licheng believe that artificial immune system has become a hot topic in the field of artificial intelligence, and related papers and research results are also increasing year by year. The international conferences of IEEE 1997 and 1998 also organized relevant thematic discussions and established the "Artificial Immune System and Application Branch". D dasgupta system analyzes the similarities and differences between artificial immune system and artificial neural network, and thinks that they are similar in the number and number of components, interaction, pattern recognition, task execution, memory learning and system robustness, but different in system distribution, communication between components and system control, and points out that natural immune system is an important source of inspiration for artificial intelligence methods. Gasper and others think that diversity is the basic feature of adaptive dynamics, and AIS is an optimization method to keep this diversity better than GA.

Due to the complexity of the immune system itself, there are few descriptions of the algorithm mechanism and few related operators. Castro, Jin Zhenjun, Du Haifeng and Jiao Lizhong. Clonal selection algorithms based on antibody clonal selection mechanism have been proposed one after another. Nohara et al. proposed a non-network artificial immune system model based on antibody unit function. At present, two more influential artificial immune network models are resource-constrained artificial immune system (RLAIS) proposed by Timmis et al. based on the concept of artificial identification ball (AR) and aiNet algorithm proposed by Leandro to simulate the response of immune network to antigen stimulation.