Simple Layout of the Starbucks1
Data Collection and Analysis1
Inter Arrival Time3
Service at the Counter4
Service Time for Barista 15
Service Time for Barista 26
Observation Table …………………………………………………………………………………………………………………………………….7
Starbucks is the largest coffee house company in the world. They have over 16,000 stores in over 50 countries. We have one of their outlets in our university. We chose to carry out our simulation project on this particular store because it would be ideal to study a system which has a queue at any time during its working hours. It would also help the company in serving their customers more efficiently and quickly, as many have limited time to waste standing in the queue. This would also give us an opportunity to use Arena in simulating the current scenario with certain assumptions, better understand the system by making changes in it and eventually to use some of the skills developed during our graduate course in Industrial Engineering to give suggestions to the company. We would have to study the present scenario like the inter-arrival time, the service time, the time in queue, the number in and the number out, the utilization, etc. and provide a solution to them, which does not involve investing on new resources, but to better utilize their current resources and serve customers quicker.
The coffee shop in our University center has one person handling the billing counter and two for preparing and serving the coffee. Our scope for the project will be to calculate the parameters which describe the functioning of the system. A simple layout of the coffee house is as shown below.
Fig.1: Simple Layout of Starbucks
Data Collection and Analysis
The data collection process essentially involved the following observations
1. Arrival Time of each Customer into the System.
2. Beginning Time for Service at Counter.
3. End Time for Service at the Counter.
4. Service Time of Barista 1.
5. Service Time of Barista 2.
6. Time of the Customer leaving the System.
The data collection process was planned before actually collecting data. It was planned to collect the time of each customer arrival. The time the customer reached the cash counter was also noted. Then, the time they left the counter was noted and finally, the time the customer left the system was noted. Also noted was the barista serving the customer, that is, whether server 1 or server 2 served the customer.
The next step was to input these values in the Input Analyzer of ARENA and observe which distribution fit best. It was essential to check if the p-value was greater than 0.05. With the first set of data collected, a distribution for the service time of barista 2 and inter arrival time could not be fit. The width of the histogram was changed and the outliers were removed as well. This did not improve the fit and the p-value was still low.
Due to this, data collection was repeated with more number of observations during the peak hour of business and the above mentioned procedure was repeated and the data collected. The distribution for the inter-arrival, service time at the counter and for the two servers were also calculated and the distribution fit, with the outliers removed and the histogram interval changed, until a high p-value was obtained. The results are given in detail below and the histograms are as shown in figure.
Simulation Model and Results:
After collecting data by observing the system and fitting the right distribution for each of the data collected, i.e., for the inter-arrival time, the service time at the counter and the service time for making and delivering the coffee, the next step was to make a simple model. The model was created using some simple modules from the basic process template. The inter-arrival time, the service time at the counter and the...