Enhancing Response Surface Methodology through Coefficient Clipping Based on Prior Knowledge

Abstract

Response surface methodology (RSM) is one of the most effective design of experiments (DoE) methods for analyzing and optimizing experiments with limited data. However, the performance of RSM is highly dependent on the quality of the experimental data (e.g., measurement error and bias). In this work, we introduce a coefficient clipping technique based on prior knowledge to address this problem in RSM. To maintain the simplicity of RSM, the representative prior knowledge of monotonically increasing/decreasing and convex/concave relationships is considered as constraints. The proposed method uses the same experimental data as typical RSM, but can more accurately analyze the relationship between the independent variable and the output response. The performance of the proposed method is verified via various case studies, including the experiment of antibiotic adsorption in wastewater.