One of the misperceptions about Monte Carlo Simulation is that it is only useful in reducing the number of “experiments” or “test runs” in conjunction with Design and Analysis of Experiments (DOE). Please don’t get me wrong it is a very powerful and useful tool for this use and is becoming more and more popular particularly during these economically challenging times.
What I’d like to touch on today, is using Monte Carlo software (@RISK) for a Lean project where you have virtually no data. When I was looking for my Lean Six Sigma Black Belt certification project, I contacted an aerospace company in the hopes they would have the need of process improvements in areas where there was much data so that I could apply all the wonderful tools I had learned about in my studies. To make a long story short, they had the need for process improvements in both manufacturing and engineering but their true needs were in the project management processes upstream of manufacturing. They regularly quoted a 12 month delivery time, but usually delivered in 17 months, with OTD rate of 22%. The only data we had to work with were, order date, date released to production and date the finished goods were shipped to the customer. The catch was, most orders were already late by the time manufacturing received all the parts to build the unit.
In order to identify the “problems”, we took one afternoon created a Process Flow Map by interviewing each of the departments and employees who were in the value stream, identified all the steps, assigned a best case, worse case and mostly likely time duration for each step, assigned distributions for each and created a Monte Carlo simulation model. The output of the model indicated a 20% OTD Rate and an average delivery rate of ~16.5 months. Not exact, but close enough to validate our model to allow the team to use @RISK’s sensitivity graphs to pinpoint which of the 80 plus process steps were contributing to the most variation. This also gave the team a starting point to hit the ground running. They are currently working on the suggested process improvements, which should save them ~$500k/year and reduce the average lead time to ~10 months.
If you would like more information, please let me know, I’d be happy to share more about the project with you.