Originally posted by: SystemAdmin

[Sylvain said:]

I would like to perform the following operations:
- do a partial search using IloSolver::solve(IloGoal goal) with a specific goal
- retrieve some information about the state of the solver after the partial search
- continue the search [u]from the current state[/u] of the solver using another specific goal

Is that possible using Ilog Solver 6.3 ?
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

[Jean-Charles Regin said:]

Hi Sylvain,

Your main issue is the use of another Goal with the current state of the search.
I think one possibility is to write your own goal who is able to call the other goals and to store the current frontier.
Do you need help for writing such a goal?
 
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

[Sylvain said:]

I need to "go out" of the search before continuing it since the choices for continuing the search cannot be made before or during the search.
Could I do something like that:

IloGoal goal1 = MyGoal1(...);
IloGoal goal2 = MyGoal2(..., &parameters);
IloSolver solver(model);
solver.startNewSearch(IloOr(goal1, goal2));
solver.next(); // First partial search
// Do something and set parameters for goal2
solver.next(); // Continue search
solver.endSearch();

#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

[Sylvain said:]

To follow up, would it be efficient to do the following:
- Perform a partial search with goal1
- Store current variable values/bounds in an IloSolution
- Continue the search with goal2 which will first restore the variable values/bounds from the IloSolution and then instantiate new variables


#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

[shaw said:]

Yes, that would work.  You could do:

solver.solve(IloRestoreSolution(env, soln) && myGoal);


Paul
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

[Sylvain said:]

Is constraint propagation performed after restoring the solution in the goal IloRestoreSolution ?
My idea is to add new bounds to the variables in a solution previously stored and perform constraint propagation in order to see the impact of this new bounds.
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

[shaw said:]

Yes, constraint propagation is carried out after the solution is restored.
#CPOptimizer
#DecisionOptimization