v9.1 no longer work for CVRP with number of stops visited per each vehicle route

[qiannathancao]

What version of OR-Tools and what language are you using?
v.9.1 vs v8.0.8283

Language: C++/Java/Python/C#
Python

Which solver are you using (e.g. CP-SAT, Routing Solver, GLOP, BOP, Gurobi)
Routing Solver: CVRP module

What operating system (Linux, Windows, ...) and version?
Windows

What did you do?
Steps to reproduce the behavior:
In order to limit vehicle stops per each route, I added the following code before setting the first solution heuristic.

    count_stop_callback = routing.RegisterUnaryTransitCallback(lambda index: 1)
    dimension_name = 'Counter'
    routing.AddDimension(count_stop_callback, 0, 45000, True, 'Counter')
    counter_dimension = routing.GetDimensionOrDie(dimension_name)
    
    for vehicle_id in range(data['num_vehicles']):
        index = routing.End(vehicle_id)
        solver = routing.solver()
        solver.Add(counter_dimension.CumulVar(index) <= data['stops'])

What did you expect to see
Expect to the number of stops per route does not exceed the stops being set i.e. data['stops'] for example 5 stops maximum

What did you see instead?
After upgrading the version to 9.0, the routing.SolveWithParameters(search_parameters) does not exit, it runs forever.
However, after re-install version 8.0.8283, it is back to normal.
Question is what has been changed could cause such confliction

Make sure you include information that can help us debug (full error message, model Proto).

Anything else we should know about your project / environment

[lperron]

Please provide a small reproducible example.

[Mizux]

0. Why not using AddConstantDimension instead ?
   

   or-tools/ortools/constraint_solver/routing.h

   Lines 554 to 570 in 2cb85b4

   	/// Creates a dimension where the transit variable is constrained to be
   	/// equal to 'value'; 'capacity' is the upper bound of the cumul variables.
   	/// 'name' is the name used to reference the dimension; this name is used to
   	/// get cumul and transit variables from the routing model.
   	/// Returns a pair consisting of an index to the registered unary transit
   	/// callback and a bool denoting whether the dimension has been created.
   	/// It is false if a dimension with the same name has already been created
   	/// (and doesn't create the new dimension but still register a new callback).
   	std::pair<int, bool> AddConstantDimensionWithSlack(
   	int64_t value, int64_t capacity, int64_t slack_max,
   	bool fix_start_cumul_to_zero, const std::string& name);
   	std::pair<int, bool> AddConstantDimension(int64_t value, int64_t capacity,
   	bool fix_start_cumul_to_zero,
   	const std::string& name) {
   	return AddConstantDimensionWithSlack(value, capacity, 0,
   	fix_start_cumul_to_zero, name);
   	}

1. I'm not sure the problem is related to this constraint (which seems legit BTW).
   Routing team did lot of change (heuristic etc) so I suspect your problem to unfortunately work better with previous implem
   seed/setup etc...

Could you provide your search parameter setup ?

Also, did you enable log search ? guided local search and time limit ?