News + Publications

Stochastic Process Simulations and Variability Analysis

In addition to knowing how inputs affect the average performance outcome in a process, it is often desirable to know how the variability of the inputs affects the overall variability of the outcome. If variability is high, then even though on average the outcome meets targets, a large percentage of the time the variability can result in performance falling below acceptable levels.

Process variability analysis can help you understand and optimize operations. It allows you to determine how often you expect to fall below targets. As well, it can help identify the inputs that are contributing the most to the outcome variability.

For one of our clients, we analyzed plant data and developed stochastic models, such as Gaussian process regression models, for each of the process inputs. Monte Carlo simulations were then used to examine how each input contributed to the final process variability. The results allowed the client to prioritize which aspects of the process should be targeted for improved control and instrumentation to get the greatest improvement in how often targets were met.