Investigation into fleet changes
Three weeks worth of historical processing schedules were considered when studying the current system. Schedules were constructed from 100 random generations and a further 20 seconds was spent searching with the simulated annealing algorithm. Each trial used the 'last plant arrival' performance measure for the schedule objective. The schedules generated by the VAS Tool had on average approximately 6 minutes of clean module stock outs per day. These stock outs force the trucks to wait for clean modules before they can depart the plant. It would seem that this downtime could be reduced if the total stock of modules was increased. A summary of the average daily performance values is provided in figure 1 below.
|Current system||Total plant idle time||Clean module stock out time||Waiting for hanging-bay||Total downtime at farms||Last plant arrival||Combined total shift time|
|week day average||34.49 min||6.50 min||65.51 min||283.21 min||630.50 min||1684.99 min|
1. Summary of the current systems performance measures.
The same three weeks of processing schedules that were used in the above section were used when investigating changes to the fleet. In total 210 trials were performed. All trials used the last plant arrival as the objective.
1. Swapping the current 26-module truck for a 22-module truck with a loader.
This change to the fleet produced the performance measures displayed in figure 2. The results include a 54% reduction in plant idle time. The change also caused farm idle time to reduce by 60%. The average maximum daily shift time of the new fleet vehicles was 30 minutes less than the current average maximum daily shift time. As expected, this was reflected by a small decrease in the average daily combined shift time.
|26 Module to 22 Module and loader||Total plant idle time||Clean module stock out time||Waiting for hanging-bay||Total downtime at farms||Last plant arrival||Combined total shift time|
|week day average||15.75 min||13.96 min||59.58 min||113.56 min||604.08 min||1679.47 min|
2. Resulting performance measures from swapping the 26-module truck
for a 22-module truck with a loader.
2. Adding a 22-module truck with a loader to the fleet so there are four trucks in total.
This addition to the fleet made no significant improvement to the schedules performance.
3. Adding a 26-module truck to the fleet so there are four trucks in total.
The addition of another 26-module truck was less attractive than the changes investigated in 1 and 2 above. The change added the least amount of value when considering the performance measures.
3. Adding to the total stock of modules
Increases to the current stock of modules were considered in isolation to estimate the effects that the current system would experience. On average, increasing the number of modules by 15 seemed to make the best improvement to the performance measure values, these are shown in figure 3. The change caused an average daily reduction of 7 minutes to plant idle time. Clean module stock outs and hanging bay delays both reduced by 30% and 40% respectively. This indicates the present inadequacy of the module stock for the current vehicle fleet. As could be expected, the resulting reduction in plant delays caused farm delays to reduce slightly as did the trucks total shift time.
|15 Module increase||Total plant idle time||Clean module stock out time||Waiting for hanging-bay||Total downtime at farms||Last plant arrival||Combined total shift time|
|week day average||27.48 min||4.35 min||37.54 min||281.98 min||621.73 min||1645.10 min|
3. Resulting performance measures from adding 15 modules to the current modules stock
4. Multiple changes
To further gauge the worth of individual fleet changes, it was necessary to consider them in conjunction with increases in the stock of modules. This was particularly important when the fleetís capacity was increased.
It was found that increasing the number of modules, whilst changing the existing 26 module truck to a 22 module truck with loader, had little to no effect on the schedules performance. This is most logically explained by the likelihood that the fleet change in itself decreases the fleetís capacity and hence the strain on the existing module stock.
Adding an additional 22-module truck to the existing fleet and increasing the module stocks showed good improvements for all three module increments (10, 15 and 20). Results suggested that plant idle time could be reduced by up to 80%, down to an average of 5 minutes a day. Other performance measure values also suggested that this particular fleet addition could achieve better measures with increases to the module stock. The average daily clean module stock out was approximately 25 minutes. This suggests an additional 30 modules added to the stock would buffer the times of clean modules stock outs. The combination of adding a 22-module truck and increasing the module stock also resulted in a decrease in the average maximum shift length of 40 minutes. The obvious downfall to such a proposed change is the extensive physical resources required. Therefore, the incurred cost of this option is likely to make it less attractive than those previously discussed.
5. Changing the shift start time
The final experiments examine the affect of filling the fleet vehicles with diesel before the beginning of the shift. Results showed that this procedure increased the average daily idle time at the plant by 12 minutes. Clean module stock outs were also increased slightly. No significant differences were found when comparing the shift times when starting at 2:20 am with starting at 2:00 am.