odtlearn.utils.callbacks

Module Contents

Classes

BendersCallback	This class contains a function that is called by the solver at
RobustBendersCallback	This class contains the function called by the solver at every node through

Functions

benders_subproblem(main_model_obj, b, p, w, i)
robust_tree_subproblem(master, i, terminal_nodes, ...)

odtlearn.utils.callbacks.benders_subproblem(main_model_obj, b, p, w, i)

class odtlearn.utils.callbacks.BendersCallback(X, obj, solver, **kwargs)

Bases: mip.ConstrsGenerator

This class contains a function that is called by the solver at every node through the branch-&-bound tree while we solve the model.

We are specifically interested at nodes where we get an integer solution for the master problem. When we get an integer solution for b and p, for every data-point we solve the sub-problem which is a minimum cut and check if g[i] <= value of sub-problem[i]. If this is violated we add the corresponding benders constraint as lazy constraint to the master problem and proceed. Whenever we have no violated constraint, it means that we have found the optimal solution.

generate_constrs(model: mip.Model, depth: int = 0, npass: int = 0)

Method called by the solver engine to generate cuts or lazy constraints.

After analyzing the contents of the solution in model variables vars, whose solution values can be queried with the x attribute, one or more constraints may be generated and added to the solver with the add_cut() method for cuts. This method can be called by the solver engine in two situations, in the first one a fractional solution is found and one or more inequalities can be generated (cutting planes) to remove this fractional solution. In the second case an integer feasible solution is found and then a new constraint can be generated (lazy constraint) to report that this integer solution is not feasible. To control when the constraint generator will be called set your ConstrsGenerator object in the attributes cuts_generator or lazy_constrs_generator (adding to both is also possible).

Args:

model(mip.Model): model for which cuts may be generated. Please note

that this model may have fewer variables than the original model due to pre-processing. If you want to generate cuts in terms of the original variables, one alternative is to query variables by their names, checking which ones remain in this pre-processed problem. In this procedure you can query model properties and add cuts (add_cut()) or lazy constraints (add_lazy_constr()), but you cannot perform other model modifications, such as add columns.

depth(int): depth of the search tree (0 is the root node) npass(int): current number of cut passes in this node

odtlearn.utils.callbacks.robust_tree_subproblem(master, i, terminal_nodes, terminal_path_dict, terminal_features_dict, terminal_assignments_dict, terminal_cutoffs_dict, initial_xi={}, initial_mins={}, initial_maxes={})

class odtlearn.utils.callbacks.RobustBendersCallback(X, obj, solver, **kwargs)

Bases: mip.ConstrsGenerator

This class contains the function called by the solver at every node through the branch-&-bound tree while we solve the model. We are specifically interested at nodes where we get an integer solution for the master problem. When we get an integer solution for b and p, for every datapoint we solve the subproblem which is a minimum cut and check if g[i] <= value of subproblem[i]. If this is violated we add the corresponding benders constraint as lazy constraint to the master problem and proceed. Whenever we have no violated constraint, it means that we have found the optimal solution.

generate_constrs(model: mip.Model, depth: int = 0, npass: int = 0)

Method called by the solver engine to generate cuts or lazy constraints.

After analyzing the contents of the solution in model variables vars, whose solution values can be queried with the x attribute, one or more constraints may be generated and added to the solver with the add_cut() method for cuts. This method can be called by the solver engine in two situations, in the first one a fractional solution is found and one or more inequalities can be generated (cutting planes) to remove this fractional solution. In the second case an integer feasible solution is found and then a new constraint can be generated (lazy constraint) to report that this integer solution is not feasible. To control when the constraint generator will be called set your ConstrsGenerator object in the attributes cuts_generator or lazy_constrs_generator (adding to both is also possible).

Args:

model(mip.Model): model for which cuts may be generated. Please note

that this model may have fewer variables than the original model due to pre-processing. If you want to generate cuts in terms of the original variables, one alternative is to query variables by their names, checking which ones remain in this pre-processed problem. In this procedure you can query model properties and add cuts (add_cut()) or lazy constraints (add_lazy_constr()), but you cannot perform other model modifications, such as add columns.

depth(int): depth of the search tree (0 is the root node) npass(int): current number of cut passes in this node