Improving Accuracy and Computational Efficiency of Optimal Design of Experiment via Greedy Backward Approach
Published in International Journal for Uncertainty Quantification, 2023
This paper introduces a backward-greedy strategy for optimal design of experiments in polynomial-chaos-based surrogate modeling. Instead of iteratively adding new samples, the method begins with a larger candidate set and removes the least informative points, improving both computational efficiency and approximation accuracy while making the choice of optimality criterion less critical in practice.
Recommended citation: Taghizadeh, M., Xiu, D., & Alemazkoor, N. (2023). "Improving Accuracy and Computational Efficiency of Optimal Design of Experiment via Greedy Backward Approach." International Journal for Uncertainty Quantification.