IFCPA: immune forgetting clonal programming algorithm for large parameter optimization problems
Maoguo Gong,Licheng Jiao,Haifeng Du,Bin Lu,Wentao Huang
DOI: https://doi.org/10.1007/11539117_116
2005-01-01
Abstract:A novel artificial immune system algorithm, Immune Forgetting Clonal Programming Algorithm (IFCPA), is put forward. The essential of the clonal selection inspired operations is producing a variation population around the antibodies according to their affinities, and then the searching area is enlarged by uniting the global and local search. With the help of immune forgetting inspired operations, the new algorithm abstract certain antibodies to a forgetting unit, and the antibodies of clonal forgetting unit do not participate in the successive immune operations. Decimal coding with limited digits makes IFCPA more convenient than other binary-coded clonal selection algorithms in large parameter optimization problems. Special mutation and recombination methods are adopted in the antibody population's evolution process of IFCPA in order to reflect the process of biological antibody gene operations more vividly. Compared with some other Evolutionary Programming algorithms such as Breeder Genetic Algorithm, IFCPA is shown to be an evolutionary strategy which has the ability for solving complex large parameter optimization problems, such as high-dimensional Function Optimizations, and has a higher convergence speed.