Design And Analyses Of Artificial Immune Algorithms

Clonal Selection Algorithms (CSAs) are a kind of important algorithms in Artificial Immune community. As an important operator of CSAs, the influence of the metadynamics strategy on the performance of CSAs has not been paid much attention. Based on Evolutionary Algorithms (EAs) and Negative Selection algorithms, Evolutionary Negative Selection Algorithms (ENSAs) have been proposed, and it can be used for combinational optimization problems and anomaly detection problems. However, the average time complexity of ENSAs for combinational optimization problems has never been studied before.In this paper, these two algorithms are studied, and the work includes,(1) Based on experiments, the influence of metadynamics on the performance of CSAs are studied and discussed. In the binary space, the traditional metadynamics usually adopts an equal probability to generate 0 and 1 for each bit of the chromosome. However, for some problems, such a metadynamics could not really increase the population diversity. In this paper, four different metadynamics strategies including the traditional metadynamics strategy and three novel metadynamics strategies are tested by four combinational optimization problems. The experimental results demonstrate these three novel metadynamics could improve the performance of CSAs for these four problems.(2) The average time complexity of ENSAs on one combinational optimization problem is analyzed. The theoretical results demonstrate that, for the Two Max function, the ENSAs with an appropriate matching threshold could perform better than the traditional (N+N) EA. Some simulation experiments on the combinational problem are also done, and the experimental results are consistent with theoretical results.(3) Furthermore, when the threshold changes, the average time complexity of ENSA is also analyzed. The theoretical results and experimental results both demonstrate that, when the threshold changes in constant, the average time complexity of ENSA is not changed.In all, three noval metadynamics are proposed in this paper, and the influence of metadynamics on the performance of CSA is analyzed. Then the average time complexity of ENSA on a combinational optimization problem is alayzed. Also, the influence of threshold on the performance of ENSA is analyzed. The results of this paper are not only important for design of Artificial Immune Algorithms for combinational optimization problem, but also for the theoretical anlyses and application of Afriticial Immune Algorithms.