Speakers and Abstracts

In determining SW Quality Readiness for customer releases we need to estimate SW Quality levels in terms of number of defects left uncovered, unresolved at the time of an intended release date. Likewise we need to estimate a likely release date when there will be acceptable quality level. There are several variable inputs to these estimates. Point Estimation is too risky if the variability in the inputs are not considered. Before launch of a Software Release, we need to estimate the expected reliability of the product when it is under actual consumer use. This helps us to estimate Warranty cost for an expected Field Failure Rate. In the situations stated above Monte Carlo simulation methods using Crystal Ball helps to determine the Software Quality Readiness Criteria and Software Reliability using multiple factors. This presentation will illustrate the use of Monte Carlo simulation related to the above situations.

In and Out of Six Sigma – Crystal Ball Translates

Scott Friesen, Lean Six Sigma Master Black Belt, Best Buy

As a Black Belt and Master Black Belt from 2005 to 2007, Scott Friesen spent a lot of his time educating his colleagues on analytic techniques to solve business problems. Pushing the edges of the existing Lean Six Sigma tool set, he began promoting the use of Crystal Ball for Monte Carlo simulation and Optimization situations. Having repatriated back into a business role within the Geek Squad services business, he has a new set of needs for the tool related to predicting business performance amongst high levels of volatility. This discussion will focus on the business applications of Crystal Ball both inside and outside of Six Sigma in the Retail industry.

Faster, Leaner Design Process with Monte Carlo Simulation

Mark Petrotta, Six Sigma Black Belt, General Dynamics

As systems become more complex and integrated, the design process itself becomes more challenging to manage. Over the past several years, the defense industry has been developing highly-integrated “Systems of Systems”, bringing together engineering teams from many companies. Design decisions made by these teams have broad implications, but are made on different schedules. With most design parameters uncertain early on, the design process tends to be iterative and inefficient.

Monte Carlo simulation offers the potential to focus the design process. The design uncertainties, when modeled using Monte Carlo techniques, provide insight into which factors are highly-leveraged, or most influential to system performance. By focusing engineering resources on resolving high-leverage decisions first, the process becomes more targeted and efficient. Monte Carlo uncertainty analysis yields a faster and leaner design process.

Optimizing Work In Process Inventory Using Monte Carlo Simulation

Arturo Valdes, Six Sigma Black Belt, TRW

Today's Automotive Marketplace demands flexibility in process and products with optimal operational costs. Optimal inventories are required to support today's environment. The use of Monte Carlo simulation is a powerful tool to optimize the cost of carrying inventory between sub-processes maximizing up-time, and evaluating current capacity vs. customer demand. The current presentation will provide an alternative to operate considering these 3 variables at the same time.

Crystal Ball for Design Engineering – Optimizing Performance and Cost

Timothy D. Williams, Six Sigma Master Black Belt, John Deere and Company

What does your customer expect from you? What is the expected life of your product? At John Deere the customer's expectation for the life of a tractor is… forever. Not too surprisingly the goal then is to optimize the product with regard to a myriad of metrics. Which begs the question "What does optimize mean?" Once that is answered, the next task is to figure out how to get there!

Six Sigma and especially Design for Six Sigma are quickly evolving. One of the most important decision-making tools in the DFSS toolbox is simulation. This presentation will put forward two projects that each address optimization from different angles: performance and cost – both essential to the business bottom line. The discussion will be clear and easy to understand, yet will challenge some common notions of what DFSS and Monte Carlo simulation are capable of delivering. A short discussion of the application of these ideas to areas other than product design will also be included.

Food for Thought: Alternative Uses for Monte Carlo Simulation in Six Sigma

Karl Luce, Master Black Belt, Oracle’s Crystal Ball GBU

Monte Carlo simulation has become indispensable in Six Sigma efforts as a virtual proving ground of design or process concepts. The Black Belt or analyst has the ability to predict, with use of transfer functions and variation of input values, the quality of a process or product CTQ before implementing production or design changes. The results can radically reduce design iteration costs and Continuous Improvement project times.

However, there are other less traditional applications for Monte Carlo analysis within the Six Sigma framework. These applications can incorporate uncertainty in teamwork solving tools (QFD, DFMEA), uncertainty in data (sample size, Gage R&R, regression samples), uncertainty in project selection decisions, and uncertainty in project cost savings. Lack of data or uncertainty in team judgments normally represent roadblocks or detours in progress. Monte Carlo can be used in these situations to frame these uncertainties and their importance to the overall project goals. Ultimately, this will direct teams to focus on the uncertainties that require more clarity and away from the uncertainties that have little impact. This presentation will provide an overview of alternative (as opposed to traditional) Monte Carlo applications in the Continuous Improvement project world.