It is perhaps best to run a virtual machine and follow the GNU/Linux instructions. But, since the software is supposed to run on several nodes of a computing cluster, MS Windows is perhaps a bad choice in the first place.
It will almost certainly be better to use the approach of
Matlab+PESTO+AMICI as mentioned earlier.
It may be possible to use Cygwin to run the software, which largely depends on the availability of the libraries we use. Cygwin is a collection of tools, some of which are usually preinstalled on a GNU/Linux system. Most importantly a C compiler like gcc and GNU make.
During installation you will have a choice of which components of
cygwin you want to install, pay special attention to the categories
Math, Science, and libraries. They contain libraries
such as the GNU scientific library
and the basic linear algebra subroutines:
BLAS. If
in doubt install everything in those categories. Make sure that gnu
make is selected for installation as well.
If you want to use vfgen to
create the C model files, you will need to compile it first, this is
the first thing the make command will do. To successfully compile
vfgen you will need
ginac and its dependeny
CLN, and also
minixml. For
parallelization, we also need any OpenMPI packages and for data
storage: hdf5.