YALMIP with GOROBI

[Zhenzhen Zheng]

Hi everyone,

I use YALMP in matlab and select the solver GOROBI:

                ops=sdpsettings('solver','gurobi');

                sol=solvesdp(F,Obj,ops)

How can I know whether the solver gets the optimal solution or not. 

Thanks, 

[Johan Löfberg]

If sol.problem = 0 Gurobi has computed an optimal solution.

[Zhenzhen Zheng]

I find that the function solvesdp returns 1 if the problem is infeasible in YALMIP Wiki. 

and I have  another question:

Can we call the function that Gurobi provide?

Thanks 

[Zhenzhen Zheng]

If we directly use the Gurobi,  the solver can return optimal state codes and the following are the state codes of Gurobi may return:

LOADED	1	Model is loaded, but no solution information is available.
OPTIMAL	2	Model was solved to optimality (subject to tolerances), and an optimal solution is available.
INFEASIBLE	3	Model was proven to be infeasible.
INF_OR_UNBD	4	Model was proven to be either infeasible or unbounded.

[Johan Löfberg]

As you maybe have noticed, I had a typo in my answer, which I fixed immeditely. As it says now, sol.problem=0 means solved to optimality within tolerances (i.e., Gurobi returned OPTIMAL)

If you want the solver specific return codes, use the option savesolveroutput

sol = solvesdp(F,obj,sdpsettings('savesolveroutput',1));
sol.solveroutput