In this blog post, we will discuss the Input Analyzer tool of the Arena Simulation program, with you. First of all, let’s talk about where we need this tool.
Let’s say we’re starting a new project. At the beginning of the project, we need to analyze the system at hand. For example, let’s create a simulation model of a call center. The durations of the phone calls operated in the call center will be different. When we create a data set on call durations, we will need the most appropriate statistical distribution that will generate new values based on this data set. At this point, we need to analyze our data. We will use the Input Analyzer tool for this exact task.
We will enter a dataset into the Input Analyzer tool and this tool will give us the most appropriate statistical distribution for this dataset. Due to this statistical distribution, we will be able to generate new values suitable for our data set. Now let’s use the Input Analyzer tool through a few examples.
Let’s consider that the customers reaching the call center are of 3 different types,
- Lost, stolen card calls
- Current customer calls
- New customer calls
We should have datasets for each customer type. The data sets that we will discuss for our sample application are as follows.
First of all, we need to export these datasets to a txt file (Picture 1). With these text files we have, we can now use the Input Analyzer tool.
After logging into the Arena program, we go to the Input Analyzer tool from the tools section. We choose “Data File – Use Existing” from the file section and the browser appears (Picture 2).
Here we select the text file we created. The Input Analyzer tool automatically analyzes the numerical values in the text file we selected and shows some results (Picture 3).
We can see the standard deviation, minimum and maximum values of our data set. Finally, we click on the fit all button at the top of the screen to determine the most appropriate statistical distribution for this data set. (Picture 4).
The Input Analyzer tool analyzed the data set we entered into it and informed us that the most appropriate statistical distribution is UNIF(1,2) (Picture 5). We can now use these distributions in our simulation model.
For example, let’s say that we want to incorporate this distribution into a process module in the arena program. We can incorporate the result we just obtained into the process module as in Picture 6.
We can analyze our data and identify the most appropriate statistical distributions, not only for call durations, but also for durations between calls or transfer durations.
In this article, we learned how to analyze our data. See you in the next article, regards.