Struct MfRbf 

pub struct MfRbf<FL, FH>
where
    FL: Fn(&[f64]) -> f64,
    FH: Fn(&[f64]) -> f64,
{
    pub low_fi: FL,
    pub high_fi: FH,
    pub options: MfRbfOptions,
    /* private fields */
}

Multi-fidelity RBF surrogate with additive correction.

Builds a correction surface δ(x) = f_H(x) - f_L(x) using RBF interpolation on a small number of high-fidelity samples, then optimizes f_L(x) + δ̂(x).

Fields

low_fi: FL

Low-fidelity model

high_fi: FH

High-fidelity model

options: MfRbfOptions

Algorithm options

Implementations

impl<FL, FH> MfRbf<FL, FH>

where FL: Fn(&[f64]) -> f64, FH: Fn(&[f64]) -> f64,

pub fn new(low_fi: FL, high_fi: FH, options: MfRbfOptions) -> Self

Create a new MfRbf optimizer

pub fn add_sample(&mut self, x: Vec<f64>)

Add a sample (evaluate both fidelities)

pub fn eval_corrected(&self, x: &[f64]) -> f64

Evaluate the corrected surrogate: f_L(x) + δ̂(x)

pub fn minimize(&mut self, x0: &[f64]) -> OptimizeResult<MultilevelResult>

Run the MFRBF optimization