A multi-objective migrating birds optimization algorithm for the hybrid flowshop rescheduling problem

In this paper, a multi-objective hybrid flowshop rescheduling problem with respect to the production efficiency and instability is addressed in a dynamic shop environment, where two types of real-time events are simultaneously considered, i.e., machine breakdown and job cancelation. To solve the problem, a multi-objective migrating birds optimization (MMBO) is proposed. In the proposed algorithm, each solution is evaluated based on the Pareto dominance relationship, and an improvement procedure is further designed to help improve the solutions quality. The fast non-dominated sorting technique is introduced to sequence the solutions in the V-shaped population. For the leader evolution, it is conducted by a Pareto-based local search method, and within the process two neighbors sets are produced to, respectively, participate in the followers evolution in the two lines. For the followers evolution, the reproduction process is introduced and the benefit mechanism is implemented by combing the genetic operators. And in the two lines, different genetic operators are employed to achieve their collaboration. For the leader change, only the promising solutions can be forwarded to the leader position. A shuffling process is proposed to help share evolutionary information between the two lines and promote their joint efforts. The performance of the MMBO is evaluated by comparing with several state-of-the-art evolutionary multi-objective optimizers, and the computational results demonstrate the effectiveness of the proposed algorithm.