TY - JOUR
T1 - Robust optimization of experimentally derived objective functions
AU - Xu, Di
AU - Albin, Susan L.
N1 - Funding Information:
Susan L. Albin is a Professor of Industrial and Systems Engineering at Rutgers University. Her research field is quality engineering and stochastic modeling. Her work has been applied in areas including semiconductor manufacturing, plastics recycling, food processing, and medical devices. Her work has been supported by NSF, FAA, DOD, Exxon and the Army. She received her doctorate from Columbia University in 1981. She was elected Secretary of INFORMS and has served on its Board of Directors. She has been an Associate Editor for Management Science and IIE Transactions, is a senior member of IIE, ASQ, and a recipient of the Exxon Education Foundation Award.
Funding Information:
The authors appreciate the helpful discussion with Dr. David Morton at University of Texas, Austin. We acknowledge the support of the NSF Industry/University Cooperative Research Center for Quality and Reliability Engineering.
PY - 2003/9
Y1 - 2003/9
N2 - In the design or improvement of systems and processes, the objective function is often a performance response surface estimated from experiments. A standard approach is to identify the levels of the design variables that optimize the estimated model. However, if the estimated model varies from the true model due to random error in the experiment, the resulting solution may be quite far from optimal. We consider all points in the confidence intervals associated with the estimated model and construct a minimax deviation model to find a robust solution that is resistant to the error in the estimated empirical model. We prove a reduction theorem to reduce the optimization model to a tractable, finite, mathematical program. The proposed approach is applied to solve for a robust order policy in an inventory problem and is compared with the canonical approach using a Monte Carlo simulation.
AB - In the design or improvement of systems and processes, the objective function is often a performance response surface estimated from experiments. A standard approach is to identify the levels of the design variables that optimize the estimated model. However, if the estimated model varies from the true model due to random error in the experiment, the resulting solution may be quite far from optimal. We consider all points in the confidence intervals associated with the estimated model and construct a minimax deviation model to find a robust solution that is resistant to the error in the estimated empirical model. We prove a reduction theorem to reduce the optimization model to a tractable, finite, mathematical program. The proposed approach is applied to solve for a robust order policy in an inventory problem and is compared with the canonical approach using a Monte Carlo simulation.
UR - http://www.scopus.com/inward/record.url?scp=0043063999&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0043063999&partnerID=8YFLogxK
U2 - 10.1080/07408170304408
DO - 10.1080/07408170304408
M3 - Article
AN - SCOPUS:0043063999
SN - 2472-5854
VL - 35
SP - 793
EP - 802
JO - AIIE Transactions (American Institute of Industrial Engineers)
JF - AIIE Transactions (American Institute of Industrial Engineers)
IS - 9
ER -