| 1 |
+++ |
| 2 |
title = "Considering uncertainties" |
| 3 |
weight = 40 |
| 4 |
+++ |
| 5 |
|
| 6 |
## Taking uncertainties into account |
| 7 |
|
| 8 |
|
| 9 |
In this example we are demonstrating how to allow for uncertainties during a Reflectometry fitting job. The sample to fit consists of twenty Titanium-Nickel bilayers. Assuming that all Titanium layers have the same thickness, the goal is to find that thickness. |
| 10 |
|
| 11 |
The reference data was generated with GENX, setting the thickness of the Ti layers equal to 3 nm. |
| 12 |
|
| 13 |
The main points to focus on here are the following: |
| 14 |
|
| 15 |
- Added artificial uncertainties to the data being fitted |
| 16 |
- Use of the the $RQ^4$ view for plotting |
| 17 |
- Use of $\chi^2$ with $L_1$ normalization as the objective metric |
| 18 |
- Setting a genetic algorithm as the minimizer |
| 19 |
|
| 20 |
|
| 21 |
Data to be fitted: {{% ref-src "testdata/specular/genx_alternating_layers.dat.gz" %}} |