Hi Geoff, Very interesting work. I think it's fair to say that the current approach for modular models in JuMP is pretty ad hoc. The goal was develop a modeling system that's similar enough to what people are used to as opposed to putting forward a new philosophy for writing down and composing models. There's no explicit concept of modularity in JuMP itself, but since we're embedded in a fully programming language, it's easy to structure your models by putting different pieces of the model in different functions. That said, I think Julia is the best platform out there at the moment for experimenting with modeling ideas and releasing them for public consumption in a usable form. With very little work you can have full access to the existing solver infrastructure, currently at 11 native interfaces and many more through writing .osil or .nl files. If you'd like to discuss more, please open up a new thread on the julia-opt list.
Best, Miles On Thursday, September 3, 2015 at 6:03:29 PM UTC-6, Geoff Leyland wrote: > > > As a way of contributing that discussion, could I point interested people > to > > > https://github.com/geoffleyland/rima/blob/master/papers/Rima-ORSNZ-presentation-2010.pdf > > It's from a while ago (before my knowledge of Julia) and now I'm really > not up-to-date with what's happening in the modular models space, but I'd > be keen to know if there's interest in the Julia community for a similar > thing in Julia, any feedback on what's there, and any pointers on how to > get started on making it Julia-like. > > Cheers, > Geoff >
