How to get request time?


Question Description:

30

1 ITM 501 Winter 2011 Term Project You will work on this project in groups of 2 or 3. The value proposition of ABC is providing not only functionality at an affordable price, but also outstanding service. As such, they always had very effective customer service, especially with their help desk that differentiated them from many other businesses at their caliber. In the early days, the owners each took a 6-hour shift staffing the help desk, which was closed from 11pm to 5am. This structure worked because ABC’s clients were all North American and from the Eastern Time zone. As the volume of business and the variety of the clientele increased, the help desk started running 24-7, and ABC hired dedicated personnel, who work 8- hour shifts. The shifts are 1: From 8 pm to 4 am, 2: from 4 am to noon, and 3: from noon to 8pm. The help desk employees are typically at the same skill and knowledge level, and they staff the shifts on a rotating basis, i.e. they all do each of the three shifts some time during the week. ABC can receive requests via phone, e-mail, or real-time (text) chat over the Web. Most of the requests are standard; therefore, the help desk personnel’s knowledge and skill levels are sufficient to handle them. In rare cases that a request needs to be processed by the development or maintenance team, the help desk employee still logs the request and then passes it on to the appropriate group. It is ABC’s policy that requests are handled on a first come-first serve basis, and a request in the queue is processed as soon as a help desk person is available. ABC does not allow their help desk staff to multi-task, because over the years they observed that they should pay individual attention to each client. Between midnight and 8am, the probabilities of an incoming request being through email, web chat, or phone are 0.4, 0.4, and 0.2 respectively. These probabilities are 0.25, 0.25, and 0.5 respectively between 8am and 4pm, and 0.3, 0.3, and 0.4 respectively between 4pm and midnight. Service time for e-mail requests is uniformly distributed between 2 and 45 minutes; service time for phone requests is normally distributed with a mean of 15 minutes and standard deviation of 8 minutes, and service time for web chat requests is given in the Appendix. 2 ABC is worried that the recent expansion of their business into new areas will result in high demand levels such that distributions of incoming service requests for different time periods will be as follows: • between 12am and 8am exponential with a rate of 20 requests per hour, • between 8am and 4pm exponential with a rate of 25 requests per hour, • between 4pm and 12am exponential with a rate of 10 requests per hour, They predict that unacceptable service levels could in turn lead to loss business either in terms of lost clientele or shrinkage of the volume of business from existing clientele. Examining data available from industry has led them to assume that the amount ($) of lost business from each minute of delay in responding to a client request (phone and web-chat) will be random with a BetaPert distribution where the best, worst, and most likely values are $0.07/min, $0.15/min, and $0.09/min respectively. The delay for these channels can be defined as the time that is elapsed between the arrival of a request and the start of a response. The cost of delay for e-mail requests (defined as the time between 30 minutes after the receipt of a service request and the time stamp on the response e-mail) is uniform between $0.03/min and $0.09/min. The costs related to help desk staffing are simply the cost of labour (help desk staff wages) and the cost of lost business. ABC pays $600/week to each of the help desk staff, each of which will be working one of the shifts every day for 7 days a week. The decision that ABC is facing is how many Help Desk staff to assign to each shift to minimize help desk related costs. Your job is to help them with this decision by simulating these costs (weekly) using Crystal Ball. You should first generate the decision alternatives and then compare the cost associated with each of these alternatives to recommend a solution. (Note that generation of good alternatives is essential to come up with a sufficiently low cost solution. Therefore, an essential part of your solution will be your strategy in the generation of these alternatives). The cost estimates should be based on 500 replications (weeks) of the simulation. 3 Deliverables (to be e-mailed directly to the professor by one of the group members (with the other group members copied) by April 5th, 2011 at 11:59pm): 1. Names and student numbers of the members of your group (not to exceed 3 members per group). Note: those of you who have not been part of any group by this date will either be assigned to a group by the professor, or will be asked to do the project by themselves. (to be submitted to Bb by April 26th, 2011 at 11:59pm, one per group): 2. The executive report (professionally written and presented in pdf or doc format) This is a report summarizing your analysis and justifying your recommendation to an ABC manager who understands their business problem, but does not necessarily know anything about simulations or anything else you have learnt in this class. 3. The analyses details (as a Crystal Ball/Excel file supported with sufficient explanations of your model) This should be intended for someone who understands how decision modeling and simulations are done. (to be e-mailed directly to the professor by each student, by April 27th, 2011 at 11:59pm) 4. Your name and a brief description of the work that each of your group members has done and your assessment (a grade out of 10) of that work 4 APPENDIX – Service Time Distribution (Web chat requests) Service Time (minutes) Probability 2 .16 3 .24 4 .21 5 .20 6 .12 8 .07

Answer

30