I recommend that we decrease the average service times for all food stations as well as reduce the minimum service time for the interactive cooking station and increase the inter-arrival time. We can implement this by offering specials a half hour to an hour before the rush period begins, have enough precooked meals available to accommodate the rush and precook the ingredients for the interactive station. We can also have the manager step in as the second cashier whenever a certain amount of time transpires.
The cashier on duty can signal to the manager by hand gestures or a light when the wait time has exceeded two minutes, respectively. I also recommend that the layout of the cafeteria be changed. The new layout will give us better organization of the lines in order to decrease confusion. Background Students have been complaining about the long wait times at the cafeteria. I observed and collected data on customer traffic in the cafeteria.
The cafeteria is most inefficient during the rush hour, 5:00pm-6:30pm, however breakfast and lunch hours do not experience the same type of rush. The long lines of the pre-cooked meals, interactive meals and cashiers have interrupted service to other areas of the cafeteria such as the drink and the salad bar stations. The data collected during the rush, located in Appendix A, indicates the entrance arrival times of each student, the service time of the cashier, precooked, and interactive lines.
The cafeteria serves about one hundred and fifty residents of Cambridge Hall and approximately one hundred residents from Nottingham Hall. The cafeteria serves hot foods, salads, snacks, sandwiches, and beverages. The data has given me information on the percentage of customers that preferred a hot meal (interactive and precooked) to snacks, the ratio of customers that prefer precooked hot meals over an interactive hot meal, line formation, service times at the different stations, arrival times and the location of the different stations.
I also learned that the peak hours of operations are from 5:00 p. m. to 6:30 p. m. and that the cafeteria has two cash registers available but only one is being utilized during the peak hours. If customers decide on a hot meal there is a 2 to 1 ratio that customers will purchase a precooked meal over an interactive meal. Through an informal customer survey, reasonable waiting times were established for the precooked line (5 minutes), the interactive line (10 minutes), and the cashier payment line (1minute).
Some of the alternatives that I will propose are to increase in the inter-arrival time, decrease the average service time, decrease the interactive service time, increase the number of cashiers, and decrease the cashier service time For the abovementioned situation, I will determine what changes to the current system are appropriate to both meet management expectations by avoiding major changes and reduce the number of complaints by improving waiting times in the most cost effective way. Analysis
To evaluate the current situation at the cafeteria, I will use simulation as a decision making model. Simulation deals with what we do not know as opposed to information that we know. Simulation is used because it allows me to test what effect a change will have on a current system before implementing the change, which will save real world time and money. The simulation model strength lies on the capability to replicate a particular situation, in our case, outcomes of various combinations of service times, arrival times and preference of food.
In addition, simulation provides better data output than theoretical models available because it uses pseudo-random numbers to repeatedly recreate input data to feed our simulated model, making outcomes more accurate since they are based on ranges rather than averages. Simulation is dependent upon all data being processed as one unit, which is done by establishing the interactions and relationships that the data has. Furthermore, since the computerized designed simulation is based on the situation at hand, situational changes can be made in a more flexible fashion making it easier to analyze them in comparison with previous scenarios.
Beginning with the data legend, the percentage of Hot Food Service (HFS), a ratio of Pre-Cooked orders to Interactive orders, the Inter-Arrival time, Cashier service time, Pre-cooked service time, Interactive service time, and the number of cashiers working are identified. The first column deals with the number of customers, which is set to simulate the probable volume of people entering the cafeteria during the 90 minute rush period. The second column is the inter-arrival times, which is generated randomly by the computer. The inter-arrival time is the time between customer arrivals.
The third column is the arrival time, which is the time between the customer arrival and the time that the previous customer arrived plus the time that passed between the two customers. The fourth column Self Service/Hot Meal divides customers into the appropriate categories based on 25% of the customers coming into the cafeteria preferring self service snacks to hot meals. The following column will take the customers who order a Hot Meal and randomly assign them to Precooked (P) or to Interactive (I) meal. The ratio of customers choosing precooked over interactive is 2 to 1.
The service column determines the service time of either precooked or interactive meals. Precooked end of service is dealt within columns seven through nine. The seventh column deals with which server the customer will go to when he or she is at the front of the line and ready to order. Each station is equipped with two servers. If a customer goes to the precooked station and server one is unavailable then he will be waited on server number two, if available. If the customer arrives at the station and both servers are busy then he will be waited on by the server that finishes first.
Column eight and nine, server one and two shows the length of time that it took for the customer to get their precooked food. Customers that choose interactive foods are shown in columns ten, eleven and twelve. These columns are set up in the same manner that the precooked end of service columns are. A separate simulation is set up for the cashier service time using similar formulas. In this spreadsheet the first column deals with the number of customers. The second and third columns are the inter-arrival and arrival times.
The third column assigns a cashier to the customer. The next column is the cashier service time which is how long the cashier takes to appropriately handle the customer. The following two columns are the end of service times for cashier one and two. These columns represent the amount of time in total the transaction takes. This simulation is run separately because the first spreadsheet is set up so that we look at each customer in the order in which they arrived. This spreadsheet reruns the arrival process and does not take into consideration the type of food purchased.
The two other spreadsheets are the food service results and cashier service results which use data tables to average to results from the simulation to get a single useful estimate for the output model. For this simulation I have found that there are many assumptions. First, I will assume that once a customer enters a waiting line they do not abandon their position. The model does not consider the fact that people waiting in the interaction line, later decide to either switch to the precooked line or go for self-service snacks and vice versa.
This simulation also assumes that customers will always be served by the next available server when in reality this is not always the case. People may not want to switch lines because they are the next in line or they anticipate problems that will cause them to wait longer. This model also assumes that people do not socialize in the dining hall when in reality dining halls are social places and a customer might see someone they know and feel like talking.
If they are talking to someone instead of waiting in line this will increase that personfs end of service time. This model does not take into account people cutting in front of other people which happens in the real world. If this happens it increases the end of service times for all the people behind that one person. This simulation also assumes an unbroken, continuous supply of food available for purchase. This means that there will never be delays by having a food service worker replenish the food inventory.
It also assumes that all electronic payment machines are fully operational with no down time. Finally the simulation model assumes that there are no customers in the cafeteria when the rush time begins. The simulation model will be run and the outputs of the data will be measured based on the manipulations that have been made to the model. This simulation does not contain any veracity errors. However, in this simulation there is a validity error occurring, the precooked to interactive ratio (2 to 1) which was causing an unreasonable wait time for students.
The original model had students waiting more then 200 minutes for interactive food which is a problem because our rush period is only 90 minutes long. The validity error will be corrected by running the base simulation with a more reasonable ratio of 5 to 1. This change in ratio improved the simulation by not having students wait unreasonable times. The wait times were decreased to about 80 minutes, while still unreasonable it is better then the original model. Using the simulation model I was able to construct outputs that identified my results as output data.
The outputs are shown in maximum and average minutes of precooked food wait time, interactive food wait time, and cashier wait times as well as the cashier idle time. I also included the percentage of customers that had to wait less then ten minutes, five minutes and more then zero minutes. This data will be what we base our decision making on when coming up with the best alternative for the cafeteria. The table below illustrates the average maximum and percent of wait times for the cafeteria at its current state.