Cafe simulation

Problem
In the simulation the goal was to find out whether it is possible to run a café with predefined spatial disposition being just one person.

To find out it was necessary to obtain data about possible ingoing flow of customers and to define activities conducted by the staff.

It was expected that the simulation should help find out, if the first hypothesis will show as faulty, the ideal/better conditions (number of personnel, number of seats, possible adjustment to a process setting etc.)

Tool
As a tool to do so the Simprocess was selected as it provides user with possibility to design processes and measure use of resources.

Spatial disposition
Spatial disposition of the Café is depicted on the image below. The designer proposes to create a space with 24 seats.

Gathering of referential data
To be able to simulate inflow of customers, it was necessary to get some referential data. Several people were asked kindly to participate in this project. Some of them did not respond, some of them did not want to share this kind of data as they probably find them sensitive. Finally, data from a place in Prague and a café in Pardubice were retrieved.

As the Prague establishment has a bit different concept, I decided not to use it directly.

Café located at city of Pardubice on the other hand was nearly perfect, as it has pretty much the same spatial proportions. Also, it is not that big agglomeration as Prague is and it is in the same region as our Café.

The number of seats at that place is 24, which is the same as the number on the first visualization of the examined Café.

Important data provided by the operational of the venue in Pardubice are following:

·      The average amount of customers per day is 80

·      There are two peaks during the day: 12:30 - 14:30 and 15:30 - 18:00

To find a measure between the café in Pardubice and our café, a metric was created.

I took number of citizens of the municipality where the café is situated and divided it by the amount of competitors plus our café which results into a value of 1565.

When done the same, but on the data from Pardubice city, the value is 2941.

The resulting ratio is 53%. With that number I multiplied the average daily traffic of the café in Pardubice which results in 43 visitors a day on average in the examined café. This amount was marked as a Realistic scenario. The 80 visitors per day was titled Optimistic scenario.

The simulation
The runtime of the simulation is set to one day – beginning at 00:00:00 and ending and 23:59:59.

Defined entities
·      Zákazník = customer

·      Working_power = auxiliary entity that helps to simulate a creation of untidiness with the direct connection to the customers

·      Dish_collecting_workingpower = also auxiliary entity that helps to simulate need for collection of the dishes. It is in a separate entity as it was necessary to make it higher priority than the regular cleaning (people would probably not like the dishes staying on the tables)

Defined resources
·      Zaměstnanec = staff member

·      Místo = a seat

Staff activities
In order to simulate the time of the visitors in the café and also the time of the staff member, it was necessary to create a list of activities that are usually executed in such an environment. To do that correctly, I involved my friends that worked/work in a café and I also utilized the fact that I personally worked in a similar establishment for few months.

We created a list of activities which we evaluated with time needed for doing those. These activities are the ones that involve staff member. Activities that are conducted only by the customers were not discussed.

Defined activities (all of a type Delay)/ used resources / duration value (minutes):
This activity is being carried out by the auxiliary entity which has priority 3 (high).

·      Collection of dirty dishes / Zaměstnanec / Nor(1.0,0.2,1)

Dining process

These activities are being carried out by the primary customer entity with high priority.

·      Making an order / Zaměstnanec, Místo / Nor(1.5,0.3,1)

·      Order deliver / Zaměstnanec, Místo / Log(5.0,1.2,45)

·      “Enjoying the perfect product” (meaning – consumption) / Místo / Nor(40.0,12.0,1)

·      Conversation with the customer / Zaměstnanec / Int(0,2)

Cleaning and maintenance

These activities are being carried out by auxiliary entity with lower priority.

·      Dishwasher manipulation / Zaměstnanec / Nor(4.0,0.5,1)

This activity batches “dirty dishes” of 10 customers.

·      Putting dishes into shelves / Zaměstnanec / (1)

·      Cleaning a table / Zaměstnanec / Nor(1.0,0.2,12)

This activity is batched on the average amount of seats by one table so it simulates cleaning of a table once there is one free.

Payment

This activity is being carried out by the primary customer entity with high priority.

·      Payment / Zaměstnanec / Nor(1.3,0.3,1)

The activity uses only Zaměstnanec as it is expected that customers will pay at the counter before they leave.

Defined entry schedule / Optimistic scenario
·      The normal - Poi(3.5,1) every 1 hour / 11:00 – 21:00

·      Peak1 - Poi(4,3) every 25 minutes / 12:30 – 14:30

·      Peak2 - Poi(4,2) every 25 minutes / 15:30 – 18:00

Defined entry schedule / Realistic scenario
·      The normal - Poi(2.1,1) every 1 hour / 11:00 – 21:00

·      Peak1 – Poi(2.5,3) every 35 minutes / 12:30 – 14:30

·      Peak2 - Poi(2.5,2) every 35 minutes / 15:30 – 18:00

Simulation results
Every simulation was done 100 times and the report was generated from the average of these simulations.

Optimistic / 1 personnel
Starting with the optimistic scenario trying to prove the one person hypothesis, we can find out that just by observing the simulation, there are customers staying in the system even after the generation of new tries stops. The last entity leaves the system slightly after 23:00, which is more than two hours after the last “customers were let it. Also, this one simulation was not that strong with just 63 entries generated. Besides this fact, that people are staying in the system that long, there was also lot of “cleaning and maintenance” to do, which the single staff person did not manage to finish before midnight.

When looking at the report, we can see that number of visitor varied between 47 and 110. Right next to it we can also see quite alarming finding that on average 16 customers remained somewhere in the system and in one case it was even 71 customers.

Another information that report gives us is a fact, that on average at one time, there was 24 customers in the system, but maximum case was 48 customers, which is the double of capacity of the café. Also, on average there was 18 people waiting for resources at one time.

All this tells us that it is not possible to run this big venue with just one person as a staff member.

Optimistic scenario with just one personnel: NO.

Optimistic / 2 personnel
The second try with 1 extra resource in staff members looks more promising. With similar numbers of customers, there is on average just 2 people left in the system. There is also on average just 2 people waiting for resources at one time, which sounds acceptable.

The average customer spends 1,5 hour in the café, which also sounds way more reasonable. With waiting time for resources with length of half an hour, there is still some space for improvement.

As we can see, the resources are far from being used to the full potential, but this might be explained by the irregular occurrence of guest and their tendency to come in groups.

The staff members spend most time on delivering the orders and preparing the food with being present for the payment as the last most time consuming activity.

Another interesting finding is the amount of time spent on different activities. On average customers spend approximately 1 hour and 12 minutes in the dining process and about 8 minutes paying.

In overall, there is still lot of space for improvement, but as a rough estimation, this simulation shows that in case the amount of customers would rise to the optimistic level, it would be necessary to incorporate two staff members.

Realistic / 1 personnel
Based on the results of previous simulations, there is an expectation that with nearly half the amount of customers coming in the café, one person should manage to take care of the venue.

With roughly 44 people visiting the café on average, there were 2 individuals left in the system by the end of the day. Maximum count of people in the café at one time is 10, but on average its 5 customers.

Customers spent usually nearly two hours in the café, with roughly 1,5 hour eating and 13 minutes paying, which leaves us with roughly 17 minutes spent on waiting for resources.

Again, the personnel is not utilized by 100%, but only 30,1%.

Overall, it seems like it is possible to run the café while being just one person in case the assumptions about amount of guests are correct. Based on the wealth of the region and on the fact, that the café is meant to be located in a quite small town, it is with high probability that the realistic scenario is more likely to be true than the optimistic one.

Propositions for improvement of the simulation
·      The generated customer should be transformed after going through the gate that simulates creation of groups of customers as it would not include the time spent waiting at the gate to the total time spent in the system, which would make the output data more clear.

·      Find a way to simulate real behavior of people sitting to the tables, as people usually do not come to a café to sit with a stranger.

·      Simulate a version with one full time customer and one extra that help would to cover only the peaks.

Practical advices resulting from the simulation
·      Hire a “cleaning woman” that helps to deal with the dishes and with the closing of the shop by the end of the day.

·      Get prepared for bigger groups of people coming in at one time – do the cleaning and other maintenance while there is not that many people in the venue.

·      In the future, when there are running costs known already, they can be included to the simulation.

Documents
Simprocess files and final reports available here:

Files here: https://drive.google.com/open?id=11rqNIemsIyRKqTONobJOpQRzQnpDlJ3N or here: