Researchers frequently use several surrogate models to address static parameter identification problems in order to represent the complex relationships between the input parameters and the corresponding output responses. What are the key considerations and challenges in selecting and combining different surrogate models to achieve optimal performance in static parameter identification? Is there a specific criterion for choosing between different models?

Does the selection and determination of the number of surrogate models depend on the type of problem or facility in static parameter identification? If so, how can one effectively select the appropriate models and determine the optimal number of models for a given problem?