When Will You Need This?
When you need to convert historical arrival data into a statistical representation of the arrivals. This especially useful for an inbound call center or a service center when the volume is greater than 100 per day, general patterns repeat weekly and more than 26 weeks of data has been collected.
The following model object will be used to create weekly pattern arrivals and import them into ProcessModel: Arrivals \ Create Weekly Pattern Arrivals.
Ease of Use: Easy
Ease of Modification: Moderate
How to Use the Model Object
1. Open Your Model
The model should have at-least one entity.
2. Insert the Model Object into Your Model
Locate the appropriate location in your model where you would like to place the model object, move the view to the location and then insert the model object.
3. Create Connections
1. Connect the arrival route that ends on the black box to an entity in your model.
2. Create a route from the black box to the rest of your model.
4. Open the Excel File
Right click on the Weekly Pattern Arrivals Sheet Excel icon and click on the third option from the bottom to open the Weekly Pattern Arrivals vx-x Excel file, where x-x represents the version number of the Excel file.
Before using this Excel file for the first time, or after changing your model name, make sure that a manual export of data is done via ProcessModel. From your model, click Tools \ Export Data, click ‘Yes’ to all prompts until the data export file opens, close the data file and then use this Excel file.
Click Enable Content if prompted
5. Enter Raw Data
Once the Weekly Pattern Arrivals vx-x Excel file opens you should land on the Input worksheet.
Copy your date and time data and paste into cell A3. You need at least 6 months of data. Use the Excel date and time format, sample data is included.
Sample data must be removed prior to entry of your data.
You must provide at-least 24 weeks worth of data
Have UK dates and time? Use the UK Dates Converter
6. Prepare Arrival Data for ProcessModel
Click the button titled Prepare Arrival Data for ProcessModel and allow 1 minute for the analysis (depends on the speed of your processor and the size of the file). Once completed you will land on the Same pattern for all days worksheet.
Follow instructions to install Histogram add-on before using the sheet
7. Choose Same or Unique Pattern for Days
After preparing your data for ProcessModel, choose one of the following two paths and then goto the relevant worksheet.
1. Same pattern for all days: When there is nominal difference in the daily pattern between day to day arrivals in the data. Daily volumes will be different.
2. Unique pattern by day of week: When there is major difference of patterns between days of the week (i.e. Sundays have a daily pattern different from Mondays).
8. Copy Data to ProcessModel
The rest of the instructions will be the same for Same pattern for all days or Unique pattern by day of week.
1. Copy the data in column B from B5 onwards by pressing the Copy Arrivals button.
2. Goto ProcessModel and open Stat::Fit via Tools \ Stat::Fit.
3. Paste the data into Stat::Fit.
4. Fit the data into a distribution by clicking Auto-fit in the toolbar and then OK.
5. Review the acceptable distributions and then click Export from the toolbar.
6. Select the acceptable distribution from the Fitted Distribution and then click OK.
7. Return to ProcessModel, click on the arrival route and then paste (CTRL + V) the distribution in the Quantity per arrival field in the Properties Dialog.
Make sure you remove the old distribution before pasting the new one in
8. Return to the Excel file and click Copy Action Logic.
9. Return to ProcessModel, click on the black box and then open the Action Logic window via the Properties Dialog. Remove the old action logic and paste (CTRL + V) in the new action logic.
The weekly patterns have now been added to your model. You can now save and then simulate the model.
a_Rand An example attribute for Weekly Pattern Arrivals Import.
How do I choose between or Same pattern for all days or Unique pattern by day of week?
You can choose between Same pattern for all days or Unique pattern by day of week by looking at the data and charts after you have prepared the data (step 6) for ProcessModel.
If there is nominal difference in the daily pattern between day to day arrivals, then use can use Same pattern for all days.
The difference between values is 1% or less
If there is a major difference in the daily pattern between day to day arrivals, then use can use Unique pattern by day of week.
The difference between values is over 7% on some days