Multi-Model Bayesian Optimization for Simulation-Based Design

Published on Nov 1, 2021in Journal of Mechanical Design2.652
· DOI :10.1115/1.4050738
Siyu Tao7
Estimated H-index: 7
(NU: Northwestern University),
Anton van Beek2
Estimated H-index: 2
(NU: Northwestern University)
+ 1 AuthorsWei Chen128
Estimated H-index: 128
(NU: Northwestern University)
Source
Abstract
References45
Newest
#1Seyede Fatemeh Ghoreishi (UMD: University of Maryland, College Park)H-Index: 10
#2Mahdi Imani (A&M: Texas A&M University)H-Index: 14
Stabilization of complex cyber-physical systems is extremely important in keeping the critical infrastructure and the environment safe. This is, in particular, critical in coupled multidisciplinary systems with several subsystems interacting with each other in an uncertain environment. The design of stabilized complex systems depends on a proper set of inputs to these subsystems, in such a way that the best stationary behavior of these systems is achieved. Despite several attempts for stabilizin...
6 CitationsSource
Nov 25, 2019 in DAC (Design Automation Conference)
#1Anton van Beek (NU: Northwestern University)H-Index: 2
#2Siyu Tao (NU: Northwestern University)H-Index: 7
Last. Wei Chen (NU: Northwestern University)H-Index: 128
view all 3 authors...
1 CitationsSource
#1Benjamin LethamH-Index: 5
#2Brian KarrerH-Index: 24
Last. Eytan BakshyH-Index: 20
view all 4 authors...
Randomized experiments are the gold standard for evaluating the effects of changes to real-world systems. Data in these tests may be difficult to collect and outcomes may have high variance, resulting in potentially large measurement error. Bayesian optimization is a promising technique for efficiently optimizing multiple continuous parameters, but existing approaches degrade in performance when the noise level is high, limiting its applicability to many randomized experiments. We derive an expr...
102 CitationsSource
Direct coupling of computer models is often difficult for computational and logistical reasons. We propose coupling computer models by linking independently developed Gaussian process emulators (GaSPs) of these models. Linked emulators are developed that are closed form, namely normally distributed with closed form predictive mean and variance functions. These are compared with a more direct emulation strategy, namely running the coupled computer models and directly emulating the system; perhaps...
11 CitationsSource
#1Siyu Tao (NU: Northwestern University)H-Index: 7
#2Kohei Shintani (NU: Northwestern University)H-Index: 4
Last. Wei Chen (NU: Northwestern University)H-Index: 128
view all 6 authors...
Enhanced collaborative optimization (ECO) is a recently developed multidisciplinary design optimization (MDO) method in the family of collaborative optimization (CO). While ECO achieves better optimization performance than its predecessors, its formulation is much more complex and incurs higher computation and communication costs, mainly due to the use of linear models of nonlocal constraints (LMNC). Consequently, ECO is often not the most desirable MDO method for large-scale and/or highly coupl...
5 CitationsSource
#1Seyede Fatemeh Ghoreishi (A&M: Texas A&M University)H-Index: 10
#2Douglas Allaire (A&M: Texas A&M University)H-Index: 18
14 CitationsSource
Aug 6, 2017 in DAC (Design Automation Conference)
#1Siyu Tao (NU: Northwestern University)H-Index: 7
#2Kohei Shintani (NU: Northwestern University)H-Index: 4
Last. Wei Chen (NU: Northwestern University)H-Index: 128
view all 7 authors...
18 CitationsSource
#1Zhen Jiang (NU: Northwestern University)H-Index: 17
#2Shishi Chen (BIT: Beijing Institute of Technology)H-Index: 4
Last. Wei Chen (NU: Northwestern University)H-Index: 128
view all 4 authors...
13 CitationsSource
#1Chen Liang (Ford Motor Company)H-Index: 7
#2Sankaran Mahadevan (Vandy: Vanderbilt University)H-Index: 75
9 CitationsSource
#1Robert B. Gramacy (U of C: University of Chicago)H-Index: 35
#2Genetha Anne Gray (SNL: Sandia National Laboratories)H-Index: 10
Last. Stefan M. Wild (Argonne National Laboratory)H-Index: 29
view all 7 authors...
Constrained blackbox optimization is a difficult problem, with most approaches coming from the mathematical programming literature. The statistical literature is sparse, especially in addressing problems with nontrivial constraints. This situation is unfortunate because statistical methods have many attractive properties: global scope, handling noisy objectives, sensitivity analysis, and so forth. To narrow that gap, we propose a combination of response surface modeling, expected improvement, an...
62 CitationsSource
Cited By2
Newest
#1Yin Liu (DUT: Dalian University of Technology)H-Index: 1
#2Li KunpengH-Index: 1
Last. Wei SunH-Index: 17
view all 6 authors...
2 CitationsSource