Require: Input parameter vector ($\mathcal{X}$)

Ensure: Parameter distribution is uniform distribution $\mathcal{Z}\sim \mathcal{U}{[0,1]}^M$

Ensure: Number of key parameters N p and total sampling set N T

Bonded design space of $\mathcal{X}$ as [lower limit, higher limit]

Number of initial sampling set using LHS, N = 10N p

Initial sampling set using LHS, N i

Initial experimental $\mathcal{X}=x_1^{(1)},\text{…},x_1^{(N)}$

corresponding original model response ${\mathcal{Y}}^{original}={\mathcal{M}}^{original}(\mathcal{X})=y_1^{(1)},\text{…},y_1^{(N)}$

Create the surrogate model ${\mathcal{M}}^{SM}$

Set threshold θ = 10−4 set threshold

if ε LOO < θ then

N new = N

${\mathcal{M}}^S$ is accepted

else if ε LOO = θ then

minimum accuracy is reached

N new = N

${\mathcal{M}}^S$ is accepted

else if ε LOO > θ then

${\mathbf{N}}^{\ast }=\underset{s\in S_N}{max}\underset{i=1,\text{…},N}{min}\parallel N_T-N_i\parallel$

find $N^{new}=N\in N_T:{\mathcal{Y}}^{SM}(N)<q$

end if