An approach to predictive-reactive scheduling of parallel machines subject to disruptions

In this paper, a new predictive-reactive approach to a parallel machine scheduling problem in the presence of uncertain disruptions is presented. The approach developed is based on generating a predictive parallel machine scheduling in such a way as to absorb the effects of possible uncertain disruptions through adding idle times to the job processing times. The uncertain disruption considered is material shortage and it is described using two parameters: number of disruption occurrences and disruption repair period. These parameters are specified imprecisely based on managerial subjective judgement and they are modelled and combined using fuzzy sets. If the impact of a disruption is too high to be absorbed by the predictive schedule, a rescheduling action is needed. In this new approach two rescheduling methods are applied, namely Left shift rescheduling and Building new schedules. The approach developed has been applied to solving a real-life scheduling problem of a pottery company. The results obtained have been satisfactory and showed the flexibility of the predictive-reactive approach.