"Johannes Pfeifer" jpfeifer@gmx.de writes:
There is a bug in getPowerDeriv. In rare cases such as entering x^(alppha-1) with x=0 in steady state and alpha=2 the derivative is wrong. The reason is that the first condition in getPowerDeriv checks whether x=0 and then sets the derivative to 0 as the derivative is p*0^(p-1) which is 0 if p~=1. But this is wrong if the exponent is p=1 as it then is 1*0^0= 1.
I pushed a fix to git where the first condition checks whether p=1. This also affects the 4.2 branch.
Thanks I pushed your fix and another commit which fixes the same issue in various places of the C++ code.
I did not backport the fix to the 4.2 branch: I have stopped backporting fixes to that branch, since there will be no more 4.2 release. We are now focused towards the release of Dynare 4.3.0.
Thanks for your contribution,