Information Technology Reference
In-Depth Information
90
80
70
60
50
40
P= 0.10, M= 10
P= 0.15, M= 10
P= 0.20, M= 10
30
0
1000
2000
3000
4000
5000
6000
Step
Fig. 4. Time-series of moving averages in 100 steps of production rate with M = 10.
The data is plotted from 100 steps.
5 Discussion
We obtained the parameter settings that lead to accidents from the results of
our experiments. In other words, the performance capacity of plant workers is
exceeded in specific plant size and plant equipment degradation rate. While
accidents occur when P =0 . 20 and M = 10, they do not occur when P =0 . 15
and M = 10. In this model, problems are generated when the probability is r×P .
Therefore, we can estimate the production rate limits from the parameters and
the results of the production rate. The worker agents can safely process events
when P =0 . 15, M = 10 and the production rate is around 70%, that is, the
worker agents can keep the plants safe when the probability is less than roughly
0.105 (0.15
0.7) when M = 10. We can calculate the production rate limit using
the probability. In this case, the production rate limits are 0.7 when P =0 . 15 and
0.525 when P =0 . 20. We can mitigate the accidents if we reduce the production
rate to make the probability less than the calculated value before the accidents
occur. The current model has no method for controlling the production rate for
such intention. We need to implement such a control method to use this model
for real world applications.
In this experiment, we varied only two parameters, and the other parameters
were fixed. Since different plants have different production rate limits, we have
to set appropriate values to model the parameters when we apply the model
to real world plants. Although real world plants have many parameters, they
×
 
Search WWH ::




Custom Search