Modified Multifidelity Surrogate Model Based on Radial Basis Function with Adaptive Scale Factor

Abstract Multifidelity surrogates (MFSs) replace computationally intensive models by synergistically combining information from different fidelity data with a significant improvement in modeling efficiency. In this paper, modified MFS (MMFS) model based on radial basis function (RBF) is proposed, which two fidelities of can be analyzed adaptively obtaining the scale factor. MMFS, an RBF was employed to establish low-fidelity model. The correlation matrix high-fidelity samples and corresponding responses were integrated into expansion determine scaling parameters. shape parameters optimized minimizing leave-one-out cross-validation error sample points. performance MMFS compared those other (MFS-RBF cooperative RBF) single-fidelity using four benchmark test functions, impacts sizes prediction accuracy also analyzed. sensitivity randomness design experiments (DoE) investigated repeating sampling plans 20 DoEs. Stress analysis steel plate presented highlight ability proposed This research proposes new multifidelity method that fully use sets, rapidly calculate parameters, exhibit good robustness.