view article

Figure 3
Schematic diagrams of the joint qNLMSA algorithm switches. Start, [{\cal P}^{({\rm start})}] = {1.70, 0.43, 1.38}, [{\cal F}\{ {\cal P}^{({\rm start})} \}] = 0.24 × 10−5, [{\bb R}_k^{\rm (start)}] = 0.51, (a) the qNLM algorithm at the end of Stage 1, [{\cal P}] = {1.49, 0.45, 1.82}, [{\cal F}\{ {\cal P} \}] = 0.68 × 10−7, [{\bb R}_k] = 0.12, (b) the SA algorithm at the end of Stage 2, [{\cal P}] = {1.60, 0.41, 2.17}, [{\cal F}\{ {\cal P} \}] = 0.79 × 10−7, [{\bb R}_k] = 0.40, (c) the qNLM algorithm at the end of Stage 1, [{\cal P}^{({\rm end})}] = {1.50, 0.50, 1.80}, [{\cal F}\{ {\cal P}^{({\rm end})}\}] = 0.34 × 10−25, [{\bb R}_k^{\rm (end)}] = 0.27 × 10−9. The magenta line depicts the behaviour of the XRDT target function [{\cal F}\{ {\cal P}\}]. The corresponding 2D cross sections at the level Z0/T = 0.5 for the 3D displacement field function [f_{\rm Ctpd}({\bf r} - {\bf r}_0)] are shown on the right-hand side.

Journal logoFOUNDATIONS
ADVANCES
ISSN: 2053-2733
Follow Acta Cryst. A
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds