Document Type : Article
Authors
1
Dept. of Industrial Engineering Tarbiat Modares University
2
Dept. of Industrial Engineering \nTarbiat Modares University
3
Dept. of Industrial Engineering\nTarbiat Modares University
Abstract
Bottlenecks are known as constraint recourses, as they govern overall system performance. Thus, bottleneck based scheduling is a useful approach. In this research, a new heuristic algorithm, based on the bottleneck, is developed for
scheduling in flexible flow shop scheduling with unrelated parallel machines. Flexible flow shop scheduling problems (FFS), with multiple unrelated machines, contain sequencing in the flow shop, where, at any stage, there exists one or
more unrelated processor. The proposed algorithm is based on the theory of constraints (TOC); a manufacturing philosophy that states: bottleneck resources govern overall system performance. Performance enhancement of the
bottleneck stage can improve overall performance, whereas, also, finding the bottleneck stage and exploiting it can optimize the whole system performance. Based on this idea, this heuristic algorithm consists of four steps: 1) finding
the bottleneck stage: a stage with the largest work load, 2) local search: in this algorithm, sequencing in the bottleneck stage is improved by applying a bottleneck-based multiple insertion procedure (BBMIP) to the initial sequence, 3) scheduling the jobs at the bottleneck stage: based on three indicators (minimum necessary time to bottleneck stage, minimum necessary time from bottleneck stage to last stage and largest processing time), this heuristic method schedules jobs in the bottleneck stage , 4) scheduling the jobs at the non-bottleneck stage; consisting of upstream stages (based on completion at a previous stage and start time in the bottleneck stage) and downstream stages (based on completion at the previous stage and minimum necessary time from the
bottleneck stage). Afterwards, finding the best position of the bottleneck in a production system can help to make a good decision for investment. In this research, the influence of bottleneck position on various performance criteria is
investigated. For this purpose, five different criteria are used: $C_{max}$ (maximum job completion time in the last stage), average job idle time, average cycle time, average lead time and average work-in process (WIP). Various experimental factors are used to design multiple scenarios to investigate the effect of the bottleneck position on performance criteria. For each scenario, 10 test problems are generated. These criteria contain bottleneck position, number of jobs, number of stages, number of machine stages, and processing time in bottleneck and non-bottleneck stages. Based on these experimental factors, 540 test problems have been generated. Results show that the bottleneck position has no significant influence on $C_{max}$ and average cycle time, but other criteria,
such as average job idle time, average lead time and average work-in process, are influenced by bottleneck position.
Keywords
)