How to get infeasible constraint using docplex python API?

Question

I am working on the Electric Vehicle Routing problem and the solution to my problem is infeasible. Now, How can I get the constraints due to which the solution is infeasible? I want to get the violated constraints in docplex using Python API.

solve_status = mdl.get_solve_status()
    if solve_status.name == 'INFEASIBLE_SOLUTION':  # or also 'INFEASIBLE_OR_UNBOUNDED_SOLUTION'
        cref = cr.ConflictRefiner()
        print('show some of the constraints that can be removed to arrive at a minimal conflict')
        cref.refine_conflict(mdl, display=True)

I tried this, but this gives me number of constraint which are violating. Not print the constraints.

Answer 1

In the docplex documentation, there are fonctions to display conflicts and iterate over them : https://ibmdecisionoptimization.github.io/docplex-doc/mp/docplex.mp.conflict_refiner.html These are docplex.mp.conflict_refiner.dislay() - Displays all conflicts. docplex.mp.conflict_refiner.iter_conflicts() - Returns an iterator on conflicts (named tuples).

Answer 2

Small example at https://github.com/AlexFleischerParis/zoodocplex/blob/master/zoorelaxationandconflict.py

from docplex.mp.model import Model
from docplex.mp.relaxer import Relaxer

from docplex.mp.conflict_refiner import ConflictRefiner

mdl = Model(name='buses')
nbbus40 = mdl.integer_var(name='nbBus40')
nbbus30 = mdl.integer_var(name='nbBus30')

mdl.add_constraint(nbbus40*40 + nbbus30*30 >= 300, 'kids')
mdl.add_constraint(nbbus40 + nbbus30 <= 7, 'maxTotalBuses')

mdl.minimize(nbbus40*500 + nbbus30*400)

mdl.solve()

mdl.report()
print(f"* solve status is: '{mdl.solve_details.status}'") #infeasible model

print()
print("------- starting relaxation")
print()

rx = Relaxer()
rx.relax(mdl)

print ("number_of_relaxations= " + str(rx.number_of_relaxations))
rx.print_information()

mdl.report()
print(f"* status after relaxation is: {mdl.solve_details.status}")
#print(mdl.solution)

print()
print("------ starting conflict refiner")
print()

cr=ConflictRefiner()
conflicts=cr.refine_conflict(mdl)
conflicts.display()


# for conflict in conflicts:
#     st = conflict.status
#     ct = conflict.element
#     label = conflict.name
#     label_type = type(conflict.element)
#     if isinstance(conflict.element, VarLbConstraintWrapper) \
#             or isinstance(conflict.element, VarUbConstraintWrapper):
#         ct = conflict.element.get_constraint()
#
#     # Print conflict information in console
#     print("Conflict involving constraint: %s" % label)
#     print(" \tfor: %s" % ct)

"""

which gives

* solve status is: 'integer infeasible'

------- starting relaxation

number_of_relaxations= 1
* number of relaxations: 1
 - relaxed: maxTotalBuses, with relaxation: 1.0
* total absolute relaxation: 1.0
* model buses solved with objective = 3800
* status after relaxation is: optimal relaxed sum of infeasibilities

------ starting conflict refiner

conflict(s): 3
  - status: Member, Variable Lower Bound: nbBus30 >= 0
  - status: Member, LinearConstraint: kids: 40nbBus40+30nbBus30 >= 300
  - status: Member, LinearConstraint: maxTotalBuses: nbBus40+nbBus30 <= 7

  """