Or-tools: Routing -Different Vehicle Different Available Time

[shashanktrivedi2018]

Hi,

I am having four vehicles, Availability of them is different and it is as per list.

horizon = [100,200,300,400]

since all vehicle has different speed, i am using -

routing.AddDimensionWithVehicleTransits(tevaluatorList,horizon,horizon,False,"Time")

Unfortunately, it is not working, however if i make horizon =300 i.e. only one value, than it works.

I am struggling in implementing different available time for vehicles. Any help is deeply appreciated.

Thanks
Shashank Trivedi

.

[Mizux]

While horizon should be the max you can still add constraint on the cumulVar of the end node of each vehicles...

solver = routing.solver()
for count, value in enumerate(horizon):
  solver.AddConstraint(routing.CumulVar(routing.End(i), dimension) < value)

[ankitgupta0310]

Hey @shashanktrivedi2018, how did you implement the different vehicle speeds? I have been struggling with the same problem, can I have an example of what you did? Any help would be nice

[shashanktrivedi2018]

While horizon should be the max you can still add constraint on the cumulVar of the end node of each vehicles...

solver = routing.solver()
for count, value in enumerate(horizon):
  solver.AddConstraint(routing.CumulVar(routing.End(i), dimension) < value)

Thanks Mizux, indeed it is great help

[shashanktrivedi2018]

Hey @shashanktrivedi2018, how did you implement the different vehicle speeds? I have been struggling with the same problem, can I have an example of what you did? Any help would be nice

    for vr in range(num_veh):       
        time_callback1 = create_time_callback1(data,data["veh_speed"][vr])        
        tevaluatorList.append(time_callback1)
add_time_window_constraints(routing, data, time_callback,horiz,tevaluatorList)
routing.AddDimensionWithVehicleTransits(tevaluatorList,300,300,False,"Time")

def create_time_callback1(data,v):
  """Creates callback to get total times between locations."""
  def service_time(node):
    """Gets the service time for the specified location."""
    return abs(data["demands"][node]) * data["time_per_demand_unit"]

  def travel_time(from_node, to_node):
    """Gets the travel times between two locations."""
    travel_time = data["distances"][from_node][to_node] / v
    return travel_time

  def time_callback(from_node, to_node):
    """Returns the total time between the two nodes"""
    serv_time = service_time(from_node)
    trav_time = travel_time(from_node, to_node)
    return serv_time + trav_time

  return time_callback

hoping it will help u.

[ankitgupta0310]

Thanks a lot for your reply @shashanktrivedi2018 ! I have been attempting a similar approach, but Im having it incorrect :/ Did you add the add_time_window_constraints() as just an additional constraint right? Can I know how you downloaded your google-or tool? Did you use the source or the binary? (https://developers.google.com/optimization/install/python/), cause I used the binary, which I think might be the root of my problem, but I'm unsure.

    tevaluator = []


    service_times = CreateServiceTimeCallback(demands, time_per_demand_unit)
    service_time_callback = service_times.ServiceTime
    for i in range(num_vehicles):
      total_times = CreateTotalTimeCallback(service_time_callback, dist_callback, vehice_speed[i])
      total_time_callback = total_times.TotalTime
      tevaluator.append(total_time_callback)

    # Note: In this case fix_start_cumul_to_zero is set to False,
    # because some vehicles start their routes after time 0, due to resource constraints.
    #add_time_window_constraints(routing,data,total_time_callback,horizon,tevaluatorList)
    fix_start_cumul_to_zero = False
    # Add a dimension for time and a limit on the total time_horizon
    routing.AddDimensionwithVehicleTransits(tevaluator,  # total time function callback
                         horizon,
                         horizon,
                         fix_start_cumul_to_zero,
                         time)

    time_dimension = routing.GetDimensionOrDie("Time")

[shashanktrivedi2018]

Please raise a new issue,it is not right to comment on the closed issue.