Littlefield Case Study

Littlefield Simulation #1: Summary

Prior to the start of the simulation, we were provided with 50 days of throughput data, which we used to forecast demand and to develop an equipment purchase plan. We used a 10-day interval average to forecast demand and then evaluated utilization data to determine points along the curve where we would require additional capacity. We determined that we would probably need three machines at Station 1 and one or two each at Stations 2 and 3. The initial forecasts did provide a good starting point, but we quickly learned the necessity for careful monitoring and near continuous (at the 24:1 simulated to real-time pace) updates. One issue we experienced was an initial misunderstanding of the Littlefield interface, which illustrated three machines at each station. Out initial utilization computations were based on starting with three machines, and while we caught this error shortly after the start of the simulation, it did cause some delay in our communication that, in retrospect, cost us some early revenue due to a delayed purchase of a second machine at Station 1. Figure 1 is the demand curve we generated prior to start of the simulation using a 10-day averaging of incoming jobs from the first 50 days with  = 3 (worksheet with utilization tables and projections available).

In retrospect, this forecast under-estimated demand, and may have caused us to delay purchase of additional machinery. Fortunately, we were also provided utilization data, which showed Station 1 had operated at capacity for 4 of the preceding 10 days prior to start of the simulation. We had identified Station 1 as the potential active bottleneck and calculated that each machine could only handle about 5-6 jobs per day.

The following provides a chronology of our actions after start of the simulation. Figure 2 is our Transaction History.

Situation: Day 54. From start of simulation at Day 50, jobs are arriving at an increasing rate. Our initial plan was to buy a 2nd machine at station 1 at about day 70. After some initial chaos in the war room, we reached consensus to accelerate our plan.

Action: Bought second machine at Station 1.

Evaluation: This was a good decision, although we should have made this purchase at Day 50, as jobs were coming in at an increasing rate and lost some revenue days 51-55.

Situation: Day 70. Jobs are queuing up at station 2 and 3 with new jobs coming in large spikes, threatening order fulfillment (and revenue) within quoted time.

Action: Changed scheduling priority at station 2 from FIFO to jobs received from station 3 (Priority on Step 4 jobs), in order to ensure final testing was not delayed on units ready for shipment to customers.

Evaluation:

...