Not expecting any, and the model is probably too complicated for
outsiders to make substantial changes. But anyone is welcome and if the request is good I see no reason why we would not try to incorporate it.
Thanks! Julia was not even in 1.0 when we started the project, but would absolutely be a great choice today. So far, only the Conopt4 solver can solve our model, which means we have to go through GAMS even if we rewrite in Julia. I saw Rackauckas' JuliaCon presentation where he talks about ML approximations and immediately got excited too.