Hi,
I think your package looks like one of the best in Julia for doing NLoc but before diving in I wanted to ask you about the potential to define vector equations. I looked through your problems in the latest version of the docs and couldn't find any equations which did matrix multiplication.
Would it be possible to define linear algebra like dx = A*x[:,j] + B*u[:,j] where x[:,j], u[:,j] are vectors of size n and A, B are matrices of size n by n?
Similarly, would it be possible to define the cost function like,
J = x[:,end]'*P*x[:,end] + sum([x[:,k]'*Q*x[:,k] + u[:,k]'*R*u[:,k] for k=1:n])
like for defining an MPC?
I understand JuMP has limits with vector valued functions, but it seems some packages have ways around this, like in NLOpt.jl.
Best,
Hi,
I think your package looks like one of the best in Julia for doing NLoc but before diving in I wanted to ask you about the potential to define vector equations. I looked through your problems in the latest version of the docs and couldn't find any equations which did matrix multiplication.
Would it be possible to define linear algebra like
dx = A*x[:,j] + B*u[:,j]wherex[:,j], u[:,j]are vectors of size n andA, Bare matrices of size n by n?Similarly, would it be possible to define the cost function like,
J = x[:,end]'*P*x[:,end] + sum([x[:,k]'*Q*x[:,k] + u[:,k]'*R*u[:,k] for k=1:n])like for defining an MPC?
I understand JuMP has limits with vector valued functions, but it seems some packages have ways around this, like in NLOpt.jl.
Best,