Re: two layer mpc and yalmip

[Johan Löfberg]

How is this any different from all the discussions we've already had.

[A_E]

difference is about making it work with time samples I mean tuning MPC in a close loop manner in a restricted time and I mean upper layer has a structure that makes it possible to have a tuning framework that tunes Q and R in an online manner not offline. it's an online auto tuning structure that is implemented by two layer optimization and it makes it possible to work with some industrial process while that offline black-box that we already talked about is too slow and needs to make MPC layer run once for all samples. here we make MPC run for each time sample and it makes a lot of difference in a chemical process etc

[Johan Löfberg]

Well as I said before, you will not be able to do the upper levels using YALMIP, and if you use some black-box solver which makes call to some simulation which uses YALMIP, we've already discussed how to do that. There is nothing more to say about that.. It is all about the details on how to do the upper level, and that is not a YALMIP issue

[A_E]

eh... excuse me professor upper level ix a yalmip that gets input u(k-1) and gives uhat(k) that goes to optimizer and also there is output y(k) from lower layer which we give to upper level and there in another output yhat(k) and we have e(k)=y(k)+yhat(k) then upper level yalmip gives another output that subtract from e(k) and goes to optimizer.

so two yalmips there and a fminsearch to act as optimizer and give tuned Q and R I suppose...