Figure 4
From: A Kriging-Based Approach to Autonomous Experimentation with Applications to X-Ray Scattering
The synthetic test functions defined in equation (12) and (13), and the corresponding error graphs depicting the agreement of the surrogate and mathematical model depending on the number of performed measurements for several different methods. Note that the two synthetic test functions are very similar, but comprise steps of different magnitudes. The difference in magnitude of these steps mainly accounts for the different results seen in the error graphs. A more prominent step will translate into a finer sampling of the space by Kriging. (a) The synthetic test functions defined in equation (12). Note the two steps. (b) The synthetic test functions defined in equation (13). Note the two large steps. (c) Graph depicting the error defined by equation (14) when approximating the synthetic test function (12), shown in (a). The data set is interpolated using linear interpolation. (d) Graph depicting the error defined by equation (14) when approximating the synthetic test function (13), shown in (b). The data set is interpolated using linear interpolation. (e) Graph depicting the error defined by equation (14) when approximating the synthetic test function (12), shown in (a). The data set is interpolated using cubic interpolation. (f) Graph depicting the error defined by equation (14) when approximating the synthetic test function (13), shown in (b). The data set is interpolated using cubic interpolation.
