Reference no: EM133137246
OR 531 Analytics and Decision Analysis - George Mason University
Simulation Optimization Project
ASSIGNMENT
Each group (2 members max) will submit a written report (2-page max) along with a simulation of the solution as an Excel file. The written report is an opportunity to practice your technical/professional writing.
BACKGROUND
An auto maintenance shop suddenly has more customers than it can handle in a given day. The shop currently has one maintenance bay and is considering adding additional bays. The current options are to continue operations with 1 bay or expand to 2, 3, or 4 bays. Each additional bay costs $10,000 per day. There is no cost for the existing bay. Each day a random number of cars arrive and the single bay serves as many cars as possible. The number of cars the bay can handle each day is also random. If there are 10 or more cars waiting to be served, the night crew is hired for $12,000 to work on cars overnight. 10 of the cars are sent to the overnight crew, all cars above 10 are sent to a competitor resulting in no cost or revenue. Any of the 10 cars the night crew cannot handle are added to the list for the next day.
Revenues are generated for each car served by the day crew and the night crew. Example: 45 cars arrive, the bay services 20 cars resulting in 25 cars waiting to be serviced. 15 of the 25 are sent away and the remaining 10 are left for the night crew. If the night crew works on 7, the final 3 cars are left for the next day.
DATA
The attached Excel file contains historical data for the auto shop for the past 60 days. Use the data supplied to determine the best distributions to use to model the system. The data includes the following:
1. Tab1: The number of cars that arrived each day for 60 days
2. Tab2: The number of cars serviced by the day crew for 60 days
3. Tab3: The number of cars the night crew serviced each night for 60 days
4. Tab4: The average revenue per car per day for 60 days
5. For the data given in tabs 1-4, your team will have to fit the distribution using analytic solver>simulation models>distributions>distribution wizard. Use the wizard to fit the data to determine the appropriate distribution to use. If given the option, use the Kolmogorov-Smirnov tab.
6. The auto shop estimates that all new bays will service cars based on a uniform distribution with a low value of 20 and a high value of 30.
7. You will have to limit the use of "psi" formulas for distributions due to the size of the simulation
a. Replace psinormal with norm.inv(rand(),mean,stdev) (input mean and stdev)
b. Replace psiuniform with randbetween(low,high) (input low and high)
GOAL
a. Use the supplied data and information above to create a simulation to find the optimal number of additional bays the company should build. The answer will be 0, 1, 2 or 3 additional bays for a total of 1, 2, 3 or 4 bays. Please include the overall profit the company would make for each of the four options and clearly identify which option your team suggests based on the optimal solutions. Your report should include charts or graphs to support your decision.
b. Using your answer in part a, evaluate how the system will perform with the number of bays your team suggests if the number of cars that arrived each day was based on a normal distribution with a mean that ranged from 100 to 125 and a standard deviation of 10. Show how the profit will change over the range of the mean for the fixed number of bays you chose in part a.
Attachment:- Simulation Optimization Project.rar