摘要

We consider the n-job, k-stage problem in a hybrid flow shop (HFS) with the objective of minimizing the maximum completion time, or makespan, which is an NP-hard problem. An immunoglobulin-based artificial immune system algorithm, called IAIS algorithm, is developed to search for the best sequence. IAIS, which is better fit the natural immune system, improves the existing AIS by the process before/after encounter with antigens. Before encounter with antigens, a new method of somatic recombination is presented; after encounter with antigens, an isotype switching is proposed. The isotype switching is a new approach in artificial immune system, and its purpose is to produce antibodies with the same protection but different function to defense the antigen. To verify IAIS, comparisons with the existing immune-based algorithms and other non-immune-based algorithms are made. Computational results show that IAIS is very competitive for the hybrid flow shop scheduling problem.