Q1: Simulating Loan Processing

ZipLoan, Inc. operates a B2B website for (relatively) speedy business loans. In this problem, you will simulate ZipLoans operation. For simplicity, we will assume that the ZipLoans operations are a continuous, 24/7 affair (for a more realistic model, this assumption could be changed by using JaamSims time series features). ZipLoans process may be modeled as follows:

1. Loan applications arrive from ZipLoans website in a memoryless manner, with an average time between applications of 1.2 hours.

2. Each loan application must be analyzed by an underwriter. The time required for underwriting is well-modeled by a triangular random variable with a minimum of 1 hour, a most likely value of 3 hours, and a maximum value of 5 hours. You keep 3 underwriters on staff at all times, and applications wait in a queue if no underwriters are available when they arrive.

3. Historically, your underwriters approve 43% of the loan applications they analyze. The remainder are rejected.

4. For each loan approved by underwriters, additional work must be done to prepare a loan offer. For each loan, the time required is well-modeled by a triangular distribution with a minimum value of 4 hours, a most likely value of 5 hours, and maximum of 8 hours. This work is performed by employees called regular staffers, and processing a loan requires one regular staffer. You have 4 regular staffers, and approved loans wait in queue if all regular staffers are currently busy.

5. After preparation, each offer is sent back to the applicant. In 75% of the cases, the applicant accepts the offer, but the remaining applicants decline (for example, because they obtained better terms from another lender). The time applicants take to decide whether to accept loan offers is well-modeled by a triangular distribution with a minimum value of 1 hour, a most likely value of 48 hours, and a maximum value of 168 hours (one week).

6. If an offer is accepted, some additional work must be done to complete the origination of the loan. This origination process takes a single regular staffer, from the same pool of 4 workers who prepare offers, an amount of time that is well modeled by a uniform distribution between 3 and 4 hours. If all regular staffers are currently busy, an offer waits in a queue until a staffer is available. Once this origination step completes the process.

Build a JaamSim model corresponding to the entire business process above. Simulate 400 days of operation, plus a warm-up time of 30 days. Then do the following:

(a) Submit a copy of your JaamSim .cfg file to Blackboard under the assignments tab, and then Homework Assignment 10, JaamSim .cfg File.

Page 2 of 3

2. (b) Answer the following questions in your submitted document, showing your work: for cases in which loans approved, use the JaamSim results to estimate how many hours pass, on average, between a customer submitting their application and receiving their loan offer (state your result to the closest 0.1 hours). What percentage of this average is queuing time? Note: the average time for a processing step can be found from the CalculatedMean or SampleMean output of its ServiceTime probability distribution.

3. (c) Answer the following questions in your submitted document, showing your work: from the time a customer accepts an offer, estimate how many hours it takes, on average, for origination of the loan to be complete (state your result to the closest 0.1 hours). Again, use the JaamSim output report and show your work. What percentage of this average is queuing time?

4. (d) Answer the following questions in your submitted document, showing your work: What percentage of the time are the underwriters busy? What percentage of the time are regular staff personnel busy?