A Methodology to Identify Physical or Computational Experiment Conditions for Uncertainty Mitigation

This paper presents a methodology to identify experiment conditions that mitigate uncertainty in complex systems. The proposed approach is based on a three-step process; 1) a sensitivity analysis is performed to identify the most influential model parameters; 2) a response surface is constructed to approximate the behavior of the system; and 3) an optimization algorithm is used to identify the experiment conditions that minimize the uncertainty in the system response.

arXiv preprint arXiv:2405.13931