Lunacek bi-Rastrigin Function: wrong documentation?
brockho opened this issue · comments
In the documentation of the bbob
functions (http://coco.lri.fr/downloads/download15.03/bbobdocfunctions.pdf), the definition of the Lunacek bi-Rastrigin Function (f24
) says that the optimum is located in $\mu_0 {\bf 1}\pm$
with $\mu_0=2.5
. In the plots, however, and in COCO's implementation, the best-parameter
value is either +1.25
or -1.25
.
Is there a mistake in the documentation or am I missing something?
In the documentation, ${\bf x}$
is multiplied by 2 in order to obtain $\hat{ {\bf x} }$
from which the $\mu_0 {\bf 1}$
is subtracted.
So the optimum in ${\bf x}$
is +1.25
or-1.25
...
We agreed on Jan 12, 2017 that the pdf should be updated in the sense that for f24, the optimum should be changed to \frac{\mu_0}{2} {\bf 1\pm}
.
see also #837
Fixed in the document source. The available pdfs will be updated timely.
The fixed document were uploaded to HAL and to http://coco.gforge.inria.fr/downloads/download16.00
and are used at the COCO gforge home and in the links at the latest old release.