Warmstarting LP

[Iain-R]

Hello all,
I was looking for assistance or guidance on warmstarting an LP with a feasible solution.
I can do this in Gurobi by setting primal variable values, as follows :

void LPWrapper::setWarmStartValues(std::vector<std::pair<int, double>>& columnStarts) {
        for (auto& startVal : columnStarts) {
            int variable = m_variables.at(startVal.first);
            double value = startVal.second;
            m_model.getVar(variable).set(GRB_DoubleAttr_PStart, value);
      }
}

Can I do something similar in COIN CLP ?
Thank you.
Iain

[jhmgoossens]

Via OsiClpSolverInterface there is WarmStart functionality available. This, however, represents a basis not just a solution.
Hope this points you in the right direction.

[Iain-R]

Thank you for your quick reply,
I am solving a series of LP problems where the solution at iteration K is a feasible solution of iteration K+1. The constraint sets may be different between iterations. With this in mind, is there a way to convert a feasible solution to a feasible starting basis?

[jhmgoossens]

Unfortunately, I'm no expert in this area, but this is what ClpModel's setColSolution or OsiClpSolverInterface's setColSolution are for. For an example, see CbcSolver3.cpp.