13. Answers to Odd Problems

Input analysis is the process of model formulation, collecting data, and fitting distributions which results in sufficient information to construct a credible simulation model.

The goal is to develop accurate decision-support information predicting how staffing decisions (factors include the number of staff or each type) would likely affect the expected waiting times of patients (key response) in the med/surg. unit as patient demand continues to increase. The “in scope” area is indicated by the oval in the flowchart below, showing the focus on med/surg. Many details about upstream and downstream units will be simplified or ignored (Fig. 13.1).

Responses of interest include monthly expected profits and expected waiting times. Controllable input variables include: One might time the interarrivals for only the rush hours (11:45–1:30 pm) and divide food into three types: hot food, other food, and drinks and record the service times for at least 20 of each type.

Assume that it takes about 60 h to gather data, make a reasonably complicated simulation model, and validate it (including iteration). This costs about $6,000 which is about as small as any professional consulting project should be to permit sustainability. (Each project typically takes significant time to recruit the business and set the terms.) It is hard for call centers generally to model the increases or avoided losses in business for their customers because of their contribution. Did customer X sign in because they heard of the great customer service? Yet, we all know intuitively that service is important for bank competitiveness. Therefore, we will focus only on potential savings relating to reduced staff expenses. Likely, the call center will be conservative about new hires and hire only as they feel overloaded (an iterative strategy). Yet, if they do purchase new software, they may need to make decisions quickly which could involve delaying or speeding up hiring processes. If through simulation, one of these processes could be trimmed by only 2 months, the simulation project would pay for itself with interest $90,000/6 = $15,000.