Closed apascualgarcia closed 4 years ago
Compatibility solved, the main issue came from the vector of times, since the stochastic version requires much longer vectors and times do not correspond any more to the index of the vector. I created a new variable (Nrealiz) which controls the number of realizations and that differ from Nrand, which controls the number of parameters generated per realization. I made several apply functions independent of the starting value of the integration time, and the lengths of the vectors are now verified and the structures generated changes accordingly.
The same script has been used to generate the new version of the stochastic model assuming total compatibility of the solver functions, which seems it is not the case. This has been started without branching, there is no risk for now because all new statements are controlled by a var called model.type.
But it should be addressed whether it is needed a new branch or script.