We have developed a simple input-output table format to describe the
desired computation that the evolved
CSNs is
supposed to carry out. Details are given in the documentation included
with the software package. With the SBMLevolver, we implemented a
hierarchical evolutionary algorithm to combine the strength of
structural network evolution with the parameter fitting capabilities
of
the
CMA-ES. The
resulting
software takes as an input a
desired computational task and tries to find - through an evolutionary
process - a
CSN (or a set of them) capable of
carrying out that task.