Originally posted by: SystemAdmin

Hi,

I want that Cplex solves my MIP by using the initial solution (a vector) that I give to it. Do you know how can I do it? What are the codes?
In fact, I have a C++ program in which I have a loop. At each iteration of the loop, I create an MIP model. I want to ask to Cplex to solve the current MIP model by using the optimal solution of the MIP model of the preceding iteration. Furthermore, the initial solution may not be feasible (to the current MIP) but I think Cplex can overcome the infeasibility. Do you how does it do that? Do I need some other codes to ask Cplex to do it or Cplex is able to detect and repair the solution automatically?

Thank you in advance,
Max.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Take a look at IloCplex::setVectors. For more details, search the forum for references to setVectors; I seem to recall some previous threads about MIP hot starts.

Paul

Mathematicians are like Frenchmen: whenever you say something to them, they translate it into their own language, and at once it is something entirely different. (Goethe)
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Using MIP starts is better than using setVectors(). You should use IloCplex::getValues() to retrieve the optimal solution at the end of an iteration and add this as MIP start using IloCplex::addMIPStart() at the beginning of the next iteration.
If you know beforehand that some of the values in the solution vector are infeasible for the next iteration it may pay off to not fix these values and provide only a partial MIP start for the next iteration.
#CPLEXOptimizers
#DecisionOptimization