A beam search algorithm to optimize robustness under random machine breakdowns and processing time variability

The vast majority of the machine scheduling research assumes complete information about the scheduling problem and a static environment in which scheduling systems operate. In practice, however, scheduling systems are subject to considerable uncertainty in dynamic environments. The ability to cope with the uncertainty in scheduling process is becoming increasingly important in today's highly dynamic and competitive business environments. In the literature, two approaches have appeared as the effective way: reactive and proactive scheduling. The objective in reactive scheduling is to revise schedules as necessary, while proactive scheduling attempts to incorporate future disruptions when generating schedules. In this paper we take a proactive scheduling approach to solve a machine scheduling problem with two sources of uncertainty: processing time variability and machine breakdowns. We define two robustness measures and develop a heuristic based on beam search methodology to optimize them. The computational results show that the proposed algorithms perform significantly better than a number of heuristics available in the literature.