Please join us in welcoming Russell Barton, Distinguished Professor of Supply Chain and Information Systems at Penn State University. He will talk about some surprisingly simple metamodels used to forecasting cycle time quantiles.
Real-time control has become increasingly difficult as manufacturing systems and their models become more complex in terms of job variety, machine flexibility and machine reliability. We consider a control policy to be a set of rules governing when jobs are released to the manufacturing floor (and perhaps to which machine in a flexible system). The objective of a control policy is to maximize efficiency (machine utilization) while meeting job demand times. A critical quantity in this planning is the quantile (say the 90% quantile) of forecast completion time for a job released to the system in its current state. We propose an analysis method based on offline simulation of a fraction of all possible states that fits a metamodel to predict the state-based cycle time quantile. The rapid calculation enabled by the metamodel allows real-time support of a dynamic strategy for job release. In several examples, a small experiment design (0.02% of all possible conditions) allows fitting a metamodel with a prediction error typically 1% or less. This work is joint with Giulia Pedrielli at Arizona State University.
Russell Barton is a Distinguished Professor of Supply Chain and Information Systems in the Smeal College of Business at Penn State and holds a courtesy appointment in the Department of Industrial and Manufacturing Engineering. From 2013-2018 he served as Senior Associate Dean for Research and Faculty. In 2017-18 he was Vice President, INFORMS Sections and Societies. From 2010-2012 he was Program Director for Service Enterprise Systems and Manufacturing Enterprise Systems at the U.S. National Science Foundation. He has a B.S. in electrical engineering from Princeton University and M.S. and Ph.D. degrees in operations research from Cornell University. He is a Fellow of IISE, a Certified Analytics Professional®, and a Senior Member of IEEE. His research has focused on the interface between applied statistics, simulation of manufacturing and service processes, and product design and manufacturing.
