How to setup inequality constraints

[tobenda]

Hey dear all, 

i am looking for a way to setup inequality constraints for the solver when using ipopt in the casadi example for the moving horizon estimation nonlinear programming problem. In the following is a code snippet of the mhe example shown online. 

----------------------------------------------------------------------------------------------------------------

# Build the multiple shooting constraints

g = []

for i in range(N-1):

  g.append( shooting["X",i+1] - phi(shooting["X",i],parameters["U",i],shooting["W",i]) )

----------------------------------------------------------------------------------------------------------------

This way, one can setup equality constraints. But what to do to setup inequality constraints?

And is there an other solver available other than the ipopt:

----------------------------------------------------------------------------------------------------------------

# Create the solver

opts = {"ipopt.print_level":0, "print_time": False, 'ipopt.max_iter':10}

nlpsol = nlpsol("nlpsol", "ipopt", nlp, opts)

----------------------------------------------------------------------------------------------------------------

Thanks in advance for your great help. 

Best regards. 

Joe

[Joris Gillis]

Dear Joe,

Inequalities can be obtained by setting the entries of lbg and ubg to nonzero values:

  lbg <= g(x) <= ubg

The bounds can be infinite if you want one-sided inequalities.

Concerning other solvers, there a quite some (seed docs.casadi.org -> Solvers, e.g. http://casadi.sourceforge.net/api/html/d4/d89/group__nlpsol.html )
However, many are commericial; we ship interfaces to them but cannot ship the solver itself.


Best regards,
  Joris