Answered

Hi, how can I launch a simulation with this distribution (aprox 55 input events each work-day during 3 months)?

I have modeled I claims management process.

Now I am trying to launch simulation. Everything is ok in a one day simulation. But I need to see a long-time picture.

How can I simulate this input event distribution?

- Aprox 55 orders per work-day

- 5 work-days per week (standard 8 hours time schedule)

- For 3 or 6 months of duration

Is it possible?

Thanks in advance

Best Answer
photo

Dear Alberto,

According to your question, Simulation works in hours. It means that if you want to simulate 5 work-days per 8 hours each day, it sums 40 hours per week. So, you need to type 40 hours in order to simulate a work-week. if you want to do for a month, you should count the days per month and multiply per hour a day (8) and so on.

whatif2

In real scenarios, processes are subjected to ever changing conditions in the availability of resources. Holidays, weekends, shifts and breaks restrict and define the true performance of a process. This calendar level predicts how a process will perform during dynamic periods of time, such as shifts, days schedules or weeks.At the end of this level you will obtain more accurate information on:

  • Sub- or over-utilization of resources.
  • Total resources costs.
  • Total activity costs.
  • Delays (time an activity waits for a resource).
  • Expected cycle time.

Here are some examples: level_4_example

Regards

Comments (4)

photo
0

Dear Alberto,

According to your question, Simulation works in hours. It means that if you want to simulate 5 work-days per 8 hours each day, it sums 40 hours per week. So, you need to type 40 hours in order to simulate a work-week. if you want to do for a month, you should count the days per month and multiply per hour a day (8) and so on.

whatif2

In real scenarios, processes are subjected to ever changing conditions in the availability of resources. Holidays, weekends, shifts and breaks restrict and define the true performance of a process. This calendar level predicts how a process will perform during dynamic periods of time, such as shifts, days schedules or weeks.At the end of this level you will obtain more accurate information on:

  • Sub- or over-utilization of resources.
  • Total resources costs.
  • Total activity costs.
  • Delays (time an activity waits for a resource).
  • Expected cycle time.

Here are some examples: level_4_example

Regards

photo
0

Thanks for your response Juan.

But I still have some doubts.

1-. How can I simulate the input comming in batchs? I mean:

- Duration = 40 h (one week, 5 work days, 8 work hours per day)

- Max imputs = 300

- Distribution I need to set up: 60 inputs (at a time) - 8h (without inputs); - 60 inputs - 8h; - 60 inputs - 8h; - 60 inputs - 8h; - 60 inputs - 8h, ....

2-. After run the simulation, if I look at the time analysis results, I do not see Lead Time results (for example, the average total time in natural days to complete process). Is there any possibility to obtain this kind of data?

Regards

photo
0

Dear Alberto,

1. To define a certain number of inputs in a time interval we would make the frequency conversion. In this case, we are required 60 inputs in 8 hours, it is a frequency of 0.1333 hours (8h / 60inputs - it must come to each system an input 0.1333 hours). This is the value entered in the field arrival int.If we are not required 60 inputs exactly, it can be used in the probability distributions in the field arrival int to model this situation. For example, a normal distribution with mean 0,1333h and standard deviation 0,001h.On the other hand, it is not possible to do the division periods to model the situation in 8 hours where any input fails.

2. The simulation results are shown in units configured in the properties of the simulation. You must select days for results to be displayed in days. Keep in mind that these units also affect the simulation parameters, so, the time duration of the tasks, arrival intervals, etc, must also be admitted to the unit defined there.Regards

photo
1

At a more basic level, I'd like to know if I can import input volumes using real-world data? In my operation we have clear cycles of peaks and troughs (eg April-June is highest, I also have smaller peaks at quarterly cycles, and off-peak periods during other months. I need to therefore define a daily volume (or at least monthly) that varies, but not based on any normal distribution curve. Is that possible?