or-toolsconstraint-programmingoperations-researchcp-sat

Model infeasible when fixing some variables in OrTools CP-SAT

I have a boolean variable in Ortools CP-SAT, let's say x. That variable is in a constraint say y = -5 * x. I have two other constraints in the model that are model.Add(y >= 0).OnlyEnforceIf(z) and model.Add(y < 0).OnlyEnforceIf(z.Not()). When I try to fix x = 1 and solve it, I am receiving as answer that the model is infeasible? Does anyone know the reason? I was expecting it to be feasible.

Solution

  • Below code snippet works out for me. I haven't specified the objective function, since I am not sure of it.

    from ortools.sat.python import cp_model as cp

model = cp.CpModel()

x = model.NewBoolVar("")
model.Add(x == 1)

z = model.NewBoolVar("")
y = model.NewIntVar(-100, 100, name = "")

model.Add(y == -5 * x)

model.Add(y >= 0).OnlyEnforceIf(z)
model.Add(y < 0).OnlyEnforceIf(z.Not())


solver = cp.CpSolver() 
# model.Minimize(y)
solver.Solve(model) # status = 4 => optimal

solver.Value(x) # x = 1
solver.Value(z) # z = 0
solver.Value(y) # y = -5