Each year, thousands of prospective students visit the University of Pittsburgh during their college decision process. One forum that is available to students already admitted to the university is the Day for Admitted Students, or DFAS. This is a program that takes place on a Saturday, and it is available several Saturdays throughout the spring. Since these DFAS involve several programs and options throughout the day, they require a good deal of planning and organization. Although the DFAS programs currently meet the needs of visitors in terms of information presented, the programs are almost always hurried. Furthermore, many people working one session of the program report that the visitors often arrive at their session late. Some report that certain presentations must be skipped due to delays and time issues. In order to evaluate the current time constraints and organization of the DFAS, a simulation model was built. The model captures each of the programs that students and families of students can attend throughout the day.
The objective of the simulation is to evaluate time constraints and organization of the Day for Admitted Students program. In order to accomplish the objective, a set of primary questions must be answered. These questions are listed in the results section. The model was used to generate statistics about each session of the program, and from there, key questions were considered, and answered using the results data. Additionally, problem areas were identified. For this analysis of problem areas and questions, see the Results section.
From the results of the simulation, it can be observed that the least waiting time for customers and highest server utilizations are observed in scenario 2. However, in practice it is not a good idea because this will lead to an extremely long, but fast moving queue. It is difficult operationally to have 8 servers with just one queue and a bad design aesthetically too. Analysis of several intermediate and practical scenarios simulated in scenarios 3 through 7 are close to the ideal situation. For the set of input data assumed for this project, Scenario 7 can be recommended as it has a reasonable value for maximum and average waiting times for a customer and the server utilization. This is the scenario with 1 lane of fast checkout on RH schedule, 4 lanes of regular checkouts on WD schedule, 1 lane of regular checkout on RH schedule and 1 lane of self checkout on WD schedule. Further, it can be observed that improving this scenario to reduce the maximum waiting time of any customer by 1.03 minutes will result in 19.32% rise in operating costs.
Results (editor's note: results are edited for brevity.)
According to the model results, visitors are late to the final destination nearly 100% of the time. The confidence interval for total time in system, for those visitors who stay for the entire program, is 5.78 hours to 7.16 hours. The scheduled time for the total program is 5.75 hours, which means that the scheduled time falls outside of the confidence interval, making it very unlikely that a visitor would go through the entire program in the allotted time. This large percentage of visitors who are late to the final destination signifies that the scheduling is not exact, and that participants cannot be expected to meet all start-times exactly. Although those running the DFAS program do make arrangements for people to arrive at the final destination later than expected, the program might benefit from a more accurate schedule that reflects this later finish time. This is true especially when the visitors are from out-of-town and may have made travel arrangements based on the finish time of the program.