Hello.
I am trying to solve Linear Assignment Problem that is actually
polynomial. The solution polyhydron must be integral. So, LP
relaxation of MIP must provide the same objective.
Right now I am solving MIP. But apparently it works very slowly. I
tried to solve the corresponding LP but the problem is that Gurobi
gives me a solution that is not a basic feasible solution but convex
combination of them, i.e. rational and not integer. I need an integer
optimal solution. I suppose method GRBModel.optimize() uses barrier
method by default. How can I switch it to simplex? It seems to me that
simplex method will give me an integral solution, right?
I found this explanation
http://www.gurobi.com/documentation/5.0/reference-manual/node694
but I can't conceive how I can apply it in my C++ environments. What
is the name of this method? What type of Gurobi variable should I use
to switch methods of optimization?
Kind regards,
Vladimir.