Response Surface Method / Response Surface Analysis

CORE DEFINITION

Response surface methodology is a statistical approach to optimization in which the response to be improved is treated as a surface curved across several factors, and a polynomial model is fitted to the measured points. Its core proposition is that a multi-factor problem can be made visible and walked, so the region of best conditions can be located by gradient instead of by luck. The method proceeds by sequential experiments, each fitted model telling the experimenter where to measure next. The key qualification is that the fitted surface only approximates the true response within the region that was actually measured.