Dynamic artificial immune system and its application to File Transfer Scheduling optimization

There are different theories and models in natural immune system, so computer science researchers have created various algorithms to simulate processes of immune system, such as immune network based models, negative selection, and clonal selection. In this paper a novel dynamic clonal selection algorithm has been used to solve File Transfer Scheduling optimization problem. In proposed algorithm, the parameters of clonal selection algorithm will be changed over generations with hope of decreasing run-time, and at the same time the performance of the algorithm remains at an acceptable level. Then after some generations a population control strategy handles the size of antibody population. Antibodies have been created such that, the degree of simultaneous sending of files be maximized for a given transfer sequence of files. This causes make-span of schedule be minimized for that sequence. The proposed algorithm has been examined on these problems with different sizes. The results of experiments show that, the rate of reaching to global optimum is acceptable.