Abstract

Rescheduling is an essential operating task to efficiently tackle uncertainties and unexpected events frequently encountered in today's complex and flexible manufacturing systems. The main purpose of this thesis is to develop a real time reactive scheduling methodology in order to respond to such disturbances and uncertainties in a cost efficient manner. In order to assess the impact of schedule changes, a compound rescheduling cost function is developed based on machine, job, and material related rescheduling activities. A Total Rescheduling (TR) approach based on the Filtered-Beam-Search-heuristic algorithm (FBS) is proposed to generate a prespecified number of cost efficient suboptimal schedules by using the proposed cost function in case of each disruption. Thereafter, the current schedule is replaced by the alternative schedule which causes the minimum rescheduling cost. Responding to each single disruption with TR may cause system nervousness and increase the operational cost. Hence, a partial rescheduling approach is developed by a Modified Filtered-Beam-Search-heuristic algorithm (MFBSR) in order to generate a prespecified number of sub optimal cost-efficient schedules with a lower rescheduling cost and fewer deviations than TR. In order to validate the performance of the proposed methodologies, TR and MFBSR, different case studies and experimental designs have been performed considering various disruption scenarios. The performance of the suggested methods in terms of rescheduling cost, makespan efficiency and stability have been compared with similar rescheduling and repair methods in the literature. The results reveal that the proposed methodologies could be considered as competitive methods in responding to disruptions in flexible manufacturing systems