A hybrid discrete artificial bee colony algorithm for multi-objective blocking lot-streaming flow shop scheduling problem

Abstract

A blocking lot-streaming flow shop (BLSFS) scheduling problem involves in splitting a job into several sublots and no capacity buffers with blocking between adjacent machines. It is of popularity in real-world applications but hard to be effectively solved in light of many constrains and complexities. Thus, the research on optimization algorithms for the BLSFS scheduling problem is relatively scarce. In view of this, we proposed a hybrid discrete artificial bee colony (HDABC) algorithm to tackle the BLSFS scheduling problem with two commonly used and conflicting criteria, i.e., makespan and earliness time. We first presented three initialization strategies to enhance the quality of the initial population, and then developed two novel crossover operators by taking full of valuable information of non-dominated solutions to enhance the capabilities of HDABC in exploration. We applied the proposed algorithm to 16 instances and compared with three previous algorithms. The experimental results show that the proposed algorithm clearly outperforms these comparative algorithms.