An Improved Dynamic NSGA-II for Unrelated Parallel Machine Rescheduling under Machine Breakdowns

We consider a rescheduling problem of the unrelated parallel machines under machine breakdowns. Rescheduling is triggered whenever machine breakdowns and repairs occur (i.e., a rescheduling point arrives). At each rescheduling point, we establish a multi-objective mathematical model where the makespan is considered as an efficiency measure and stability is gauged in terms of the sum of deviation in completion times of jobs in the previous and new schedule. Then, we develop an improved dynamic NSGA-II to fast-track non-dominated solutions, where a population re-initialization strategy is used to effectively respond to machine breakdowns and repairs. Extensive experiments were conducted to demonstrate the performance of the proposed algorithm.