As we are concerned primarily with illustrating local search in this example, we do not spend much effort in finding a good first solution. It should be noted however that in many cases, finding a good first solution is often important for good local search performance. We use the IloRankForward predefined goal and take the first feasible solution that is found.

  IloGoal g = IloRankForward(env) && IloInstantiate(env, costVar);
  solver.startNewSearch(g);

  if(solver.next()) {
    IlcInt best = (IlcInt)solver.getValue(costVar);
    solver.out() << "Solution for Cost " <<  best << endl;
    solver.printInformation();
    globalSolution.store(scheduler);
    lsSolution.store(scheduler);
  }

After finding the solution, we call store on the lsSolution so that we are then ready to begin local search.