Initial operations strategy
Prior to the commencement of the simulation, we examined the 50 days of historical data to glean as much information as we could about the operations. We performed some analysis in Excel and created a dashboard to illustrate various data. Specifically, we regressed the prior 50 days of jobs accepted to forecast demand over the next 2 - 3 months within a 95% confidence interval. The yellow and grey lines represent the maximum and minimum variability, respectively, based on two standard deviations (95%).
Exhibit 1: Forecasted and actual demand by Day 50 and Day 270
Our two primary goals at the beginning of the simulation were as follows:
1) Eliminate bottlenecks and increase capacity in order to meet forecasted demand
2) Decrease lead time to 0.25 days in order to satisfy Contract 2 and maximize revenue
In order to achieve these goals, we would need to know the capacity and throughput time of the entire system. We used the time required by each machine to process a lot to calculate capacity per station and then capacity of the entire production line (380 kits/day or 6 orders/day). In Exhibit 2 we can observe that the capacity of the production line is given by the station that produces the least number of units per day.

Exhibit 2: Capacity of the production line
The Decisions
We decided to work with the maximum variability of demand because there was a penalty for late jobs and because there was no revenue for orders that took more than “x” days. Since the simulator runs at a rate of 34 simulated days per each real day, we had to estimate the demand for the following 34 days.
Day 1
We identified the bottleneck as the step with the least capacity, keeping in mind that steps 2 and 4 together would determine the capacity of the tester. Our strategy was to buy a machine for the step identified as the bottleneck, and then reassess