Bottleneck-based heuristic for three-machine flow shop scheduling
This paper considers a 3 machine flow shop (M1M2M3) with tendency of dominant (bottleneck) machine at M1. The developed bottleneck based heuristics from previous studies are considered in this case by Hezzeril (2010) and Irwan (2010), but were only tested for dominant machine at M2 and M3 respect...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/2724/1/24p%20MOHD%20SALLEH%20ABDUL%20RAHIM.pdf http://eprints.uthm.edu.my/2724/2/MOHD%20SALLEH%20ABDUL%20RAHIM%20COPYRIGHT%20DECLARATION.pdf http://eprints.uthm.edu.my/2724/3/MOHD%20SALLEH%20ABDUL%20RAHIM%20WATERMARK.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This paper considers a 3 machine flow shop (M1M2M3) with tendency of dominant
(bottleneck) machine at M1. The developed bottleneck based heuristics from previous
studies are considered in this case by Hezzeril (2010) and Irwan (2010), but were only
tested for dominant machine at M2 and M3 respectively. The heuristics have
successfully produced 67.24% of optimum solution at the middle process or M2 and
90.80% at the last process or M3 for 6 jobs problem. While for 10 jobs problem, the
heuristics can produce 14.64% at M2 and 90.98% at M3. As an extension of this study,
the bottleneck based heuristic scope is enlarged by developing a new heuristic for
dominant machine at M1 and combining it with the previously developed heuristics for
dominant machine at M2 and M3. The main objective is to develop scheduling heuristic
to evaluate the performance at M1 based on bottleneck analysis for M1M2M3 flow shop
and to combine with the developed heuristics from previous studies. The computer
program involved were Microsoft Excel and Visual Basic for Applications (VBA) and
the test of performance were conducted at 6 and 10 jobs problem. A simulated random
data within specific limitation being assigned at each job’s processing time produces
new recommended job arrangements. The generated makespan was compared with
optimum makespan from complete enumeration and lower bound (LB) analysis. Total
sets of 1000 simulated data at 6 and 10 jobs were allocated into 3 dominance level of
P1DL; weak, medium, and strong. Optimal solutions were obtained based on the total
results data that produce the ratio of 1. Based on the results, 62.40% of the solution
generated is optimum result for 6 jobs while 56.33% of the solution generated equals to
lower bound for 10 jobs. The heuristic performed moderately and decreased slightly
when number of jobs increased, showing that BMM1 heuristic is more suitable for lesser
number of jobs. |
|---|