Bottleneck-based heuristic for permutation flowshop scheduling

• Main Author: Isa, Noor Amira
• Format: Thesis
• Language: English; English; English
• Published: 2017
• Subjects: QA Mathematics; QA150-272.5 Algebra
• Online Access: http://eprints.uthm.edu.my/837/1/24p%20NOOR%20AMIRA%20ISA.pdf; http://eprints.uthm.edu.my/837/2/NOOR%20AMIRA%20ISA%20COPYRIGHT%20DECLARATION.pdf; http://eprints.uthm.edu.my/837/3/NOOR%20AMIRA%20ISA%20WATERMARK.pdf; http://eprints.uthm.edu.my/837/
• Institution: Universiti Tun Hussein Onn Malaysia

The newly heuristic is developed by introducing the bottleneck-based concept that was applied at the beginning of initial solution determination. The heuristic is known as Bottleneck-Based (BNB) heuristic. The previous study found that the scheduling activity become complex when dealing with large number of machine, m > 2, it is NP-hardness. Thus, the main objective of this study are to propose and develop a new heuristic for solving permutation flowshop scheduling by considering four-machines and n-jobs (n = 6, 10, 15, 20). Three phases were applied into this study in order to boost the makespan performance of the proposed heuristic. Two phases (bottleneck identification phase and initial sequence arrangement phase) were newly introduced and combined with the existing famous Nawaz, Enscore and Ham (NEH) insertion technique. There are four potential dominant machines (M1, M2, M3, M4) clustered as bottleneck machines. A total 1000 set random processing time for each job sizes was tested using Excel simple programming with built in Visual Basic for Application (VBA). The heuristic performance was evaluated based on the average makespan ratio, average percentage error, and percentage of solution performance obtained. This study considered the NEH heuristic as the best and appropriate tool for comparison purpose since NEH heuristic is the best performing heuristic in minimizing the makespan. The heuristic evaluation result showed that the BNB heuristic is performing better than the NEH at bottleneck machines M1, M3 and M4. While, heuristic verification result showed that the bottleneck algorithms performed the best in minimizing the makespan for set of problems with bottleneck machine M4. However, the overall result showed fluctuate values over the size of jobs. The result of this study shown that the developed BNB heuristic achieved good performance in solving small sized problems however further modification is needed for medium and large sized problems.