Open Lines with Finite Buffers - Diagnostic
The practical worst case can still be used as an internal benchmark for open lines with finite buffers. That is, if the performance of the line falls below the practical worst case performance, then this is an indication that there is opportunity for improvement. In addition to the operational improvements suggested for the closed line, an option for improving the line may be the addition of buffer space. It may be that the line could be made to perform above the PWC benchmark by adding buffer space at the appropriate location, or one may have to run the line at a lower release rate, as shown in the example below. The practical worst case throughput may be obtained from the diagnostic tool as before.
Example
Consider the example line we
saw before while illustrating the principle. The data is reproduced here
for convenience.
| Station | Processing Time | Mean Time to Repair | Mean Time to Failure |
| Station 1 | 1.0 | 1.0 | 5.0 |
| Station 2 | 1.0 | 1.0 | 5.0 |
| Station 3 | 1.2 | 1.0 | 5.0 |
| Station 4 | 1.0 | 1.0 | 5.0 |
The bottleneck rate for this line = 1/1.2 jobs/minute and the raw processing time = 1 + 1 + 1.2 + 1 = 4.2 minutes. If we run this system at a release rate of 1/1.5 = 0.67 jobs per minute, then even with eight buffers between each station, the WIP of the line is around 12 jobs (see simulation output). Using the diagnostic tool, we find the practical worst case throughput of this line to be 0.69 jobs/minute. Thus, as suggested by the diagnostic tool, significant improvement opportunities may exist.
Along with the other improvement opportunities
that exist, we could also reduce the release rate. For example as suggested
by the simulation output, at the release rate of 1/0.75 = 1.33 jobs/minute,
the WIP drops to around 4 jobs. At this WIP level, the PWC throughput may
be calculated as 0.51 jobs/minute using the diagnostic tool and hence we
are above the practical worst case performance. In fact as may be verified
using the same procedure above, the line continues to be worse than the
PWC with a release rate of 1/1.5 jobs/minute, and better than the PWC at
release rate of 1/0.75 jobs/minute for all buffer sizes between the stations.
|
|
|