Original Date: 06/05/2006
Revision Date: / /
Information : Monte Carlo Simulation for Risk Analysis
Rockwell Collins’ Advanced Manufacturing Technology is undertaking the use of a new process using Monte Carlo simulation to gain general insight into the nature of processes, identify problems with system designs, and manage risk by understanding costs and benefits. The introduction of this new capability will overcome the fundamental limitations of standard spreadsheet analysis.
Rockwell Collins has initiated the use of Monte Carlo simulation to measure the effects of uncertainty. Historically, Rockwell Collins has used traditional spreadsheet analysis that attempted to identify uncertainty in one of three ways: Point estimates (using only the most likely values for the uncertain variables)
Range estimates (typically using best-case, worst-case, and most likely variables)
Traditional spreadsheet analysis, however, exhibits the fundamental limitations of ordinary spreadsheets. For example, the user can only change one cell at a time, preventing the exploration of an entire range of outcomes. What-if analysis always results in a single-point estimate that does not include the likelihood of achieving any particular outcome. While single-point estimates might depict what is possible, they do not depict what is probable.
The use of random numbers to measure the effects of uncertainty have made Monte Carlo simulation beneficial to Rockwell Collins, with features that include: Acquisition of general insight into the nature of the process
Identification of system design problems
Understandable costs and benefits are understandable that contribute to risk management
User-friendly and flexible spreadsheet depiction
Rockwell Collins demonstrated two specific examples of Monte Carlo simulation that include a design-to-cost (DTC) material-estimating model, which is a future application under consideration, and a liquid crystal display (LCD) yield-analysis model. The DTC example depicts a representative range and associated probabilities for a material cost estimate. The model’s forecast chart shows the entire range and associated probability for an assembly given the various sources of the cost estimates. The model’s combination of cost and source of estimate depicted which part numbers contributed the most overall variation. The benefits include a more comprehensive analysis of underlying uncertainties while contributing to sensitivity analysis of parts that might need further consideration. The LCD yield analysis was undertaken to determine the feasibility of an accurate prediction of scrap reduction costs if two noncritical design specifications on LCD bubble defects were modified.
Both application demonstrated benefits to Rockwell Collins that include the description of possible range values for each uncertain cell in a spreadsheet, easily displayed forecast charts that show an entire range of possible outcomes and the likelihood of achieving them, easy capture of scenario results, display of critical components, and inexpensive evaluation of decisions prior to implementation.
For more information see the
Point of Contact for this survey.