This instruction aims to reproduce the results in the paper “Stacking designs: designing multifidelity computer experiments with target predictive accuracy” by Sung, Ji, Mak, Wang, and Tang (2024), SIAM/ASA Journal on Uncertainty Quantification, 12(1), 157-181.

The code folder reproduces the result in Section 5 of the manuscript.

• currin_example.R reproduces the results of Section 5.1.
• poisson_example.R reproduces the results of Section 5.2.
• blade_example.R reproduces the results of Section 5.3.

The following R packages are required to run the code:

• matlabr (>=1.5.2)
• randtoolbox (>=2.0.2)
• RColorBrewer (>=1.1-3)
• ggplot2 (>=3.3.6)
• gridExtra (>=2.3)
• plgp (>=1.1-11)

MATLAB needs to be installed in order to run the finite element simulations. The MATLAB code can be accessed from the folder Rmatlab_files.