AR1 co-kriging with Bayesian infill outperforms multi-fidelity neural networks with non-Bayesian infill on an outlet guide vane problem under a 20-sample high-fidelity budget, and POD dimension reduction improves every strategy.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.flu-dyn 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Bayesian and non-Bayesian multi-fidelity surrogate models for multi-objective aerodynamic optimization under extreme cost imbalance
AR1 co-kriging with Bayesian infill outperforms multi-fidelity neural networks with non-Bayesian infill on an outlet guide vane problem under a 20-sample high-fidelity budget, and POD dimension reduction improves every strategy.