add symmetry documentation in header

docs/gx_ac.md

@@ -36,7 +36,7 @@ g_p(z_i) = \begin{dcases}  f(z_i) & p=1 \\\;\\ \frac{g_{p-1}(z_{p-1})-g_{p-1}(z_
 ```
 
 Padé approximants are known to be [numerical instable](https://doi.org/10.1093/imamat/25.3.267). The GX-AC component uses two strategies to numerically stabilize the interpolation. 
-First, it incorporates a [greedy algorithm for Thiele Padé approximants](https://pubs.acs.org/doi/full/10.1021/acs.jctc.3c00555) that minimizes the numerical error by reordering of the reference points. A validation of the greedy algorithm can be found in this [reference](https://pubs.acs.org/doi/full/10.1021/acs.jctc.3c00555). Additionally it is possible to use the component with a  higher internal numerical floating point precision. This helps reducing the numerical noise caused by [catastrophic cancellation](https://doi.org/10.1145/103162.103163). Catastrophic cancellation occurs when rounding errors are amplified through the subtraction of rounded numbers, such as double-precision floating-point numbers commonly used in most programs. This is implemented using the [GNU Multiple Precision (GMP) library](https://gmplib.org/) which allows floating-point operations with customizable precision. Furthermore, it is possible to impose various symmetries onto the Padé model using the GX-AC component. To maximize performance, the evaluation of the Padé  model uses the [Wallis algorithm](https://numerical.recipes/book.html), which minimizes the number of divisions, an operation that is computationally expensive, especially for complex floating-point numbers and even more so for higher-precision complex numbers.
+First, it incorporates a [greedy algorithm for Thiele Padé approximants](https://pubs.acs.org/doi/full/10.1021/acs.jctc.3c00555) that minimizes the numerical error by reordering of the reference points. A validation of the greedy algorithm can be found in this [reference](https://pubs.acs.org/doi/full/10.1021/acs.jctc.3c00555). Additionally it is possible to use the component with a  higher internal numerical floating point precision. This helps reducing the numerical noise caused by [catastrophic cancellation](https://doi.org/10.1145/103162.103163). Catastrophic cancellation occurs when rounding errors are amplified through the subtraction of rounded numbers, such as double-precision floating-point numbers commonly used in most programs. This is implemented using the [GNU Multiple Precision (GMP) library](https://gmplib.org/) which allows floating-point operations with customizable precision. Furthermore, it is possible to impose various symmetries onto the Padé model using the GX-AC component. To maximize performance, the evaluation of the Padé  model uses the [Wallis algorithm](https://numerical.recipes/book.html), which minimizes the number of divisions, an operation that is computationally expensive, especially for complex floating-point numbers and even more so for higher-precision complex numbers. The component also allows symmetry-constraint Padé models, for a full list of supported symmetries see the section "Usage". 
 
 
 # Benchmarks