Multi-population Adaptive Evolutional Algorithm Based on Immune Theory

On the basis of multi-population evolution, immune theory is combined with evolution algorithm, selection, memory, clone, hyper-mutation and concentration control operators are defined, where memory operator is used to memory good gene information of parents and supervise creation of offspring, adaptive hyper-mutation operator enables individuals to confirm the search field according to self choiceness degree and number of evolution generations and concentration control operator is applied for diversity of groups. And then a novel immune algorithm based on multi-population. Emulational results show that IAM has efficient convergent speed and can be converged to the global optimal point. Compared with multi-population genetic algorithm and mind evolutionary computing, its convergent speed and convergent probability are larger