Total load greater than truck capacity

[AndresJ89]

Hello,

I'm trying to solve a VRP with 6 depots (i1 to i6), and 16 customers (k1 to k16). Each vehicle starts and returns to the same depot. The trucks have the following capacity [15, 100, 100, 100, 100, 100]. Below is the solution I'm getting. Notice that the capacity constraint holds for all the routes EXCEPT for vehicle 2 for which the final load is ~101.43. My plan is to later feed this solution to a problem in GAMS, however, this violation of the capacity results in infeasibilities in GAMS.

Route for vehicle 1:

 i1 Load(0) ->  i1 Load(0)

Distance of the route: 0.0m

Load of the route: 0

Route for vehicle 2:

 i2 Load(0) ->  k1 Load(19.642722117202254) ->  k5 Load(39.2854442344045) ->  k7 Load(59.141776937618005) ->  k4 Load(81.78449905482026) ->  k2 Load(101.42722117202251) ->  i2 Load(101.42722117202251)

Distance of the route: 639039.0m

Load of the route: 101.42722117202251

Route for vehicle 3:

 i3 Load(0) ->  i3 Load(0)

Distance of the route: 0.0m

Load of the route: 0

Route for vehicle 4:

 i4 Load(0) ->  i4 Load(0)

Distance of the route: 0.0m

Load of the route: 0

Route for vehicle 5:

 i5 Load(0) ->  k16 Load(21.429111531190987) ->  k14 Load(41.07183364839324) ->  k12 Load(60.71455576559549) ->  k6 Load(80.35727788279775) ->  k13 Load(100.0) ->  i5 Load(100.0)

Distance of the route: 162355.0m

Load of the route: 100.0

Route for vehicle 6:

 i6 Load(0) ->  k9 Load(19.642722117202254) ->  k11 Load(41.07183364839325) ->  k10 Load(60.7145557655955) ->  k15 Load(80.35727788279775) ->  k8 Load(100.0) ->  i6 Load(100.0)

Distance of the route: 571027.0m

Load of the route: 100.0

Total distance of all routes: 1372421.0m

Total load of all routes: 301.42722117202254

This is how I'm defining the truck capacity constraint:

    routing.AddDimensionWithVehicleCapacity(

        demand_callback_index,

        0,  # null capacity slack

        data['vehicle_capacities'],  # vehicle maximum capacities

        True,  # start cumul to zero

        'Capacity')

Any suggestions are appreciated.

Andres J.

[Laurent Perron]

The solver is integral only, and python silently round all floating point values to the nearest int. 

Make sure you only use integer value, everywhere. 

--
You received this message because you are subscribed to the Google Groups "or-tools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to or-tools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/or-tools-discuss/664f947f-1803-4d5c-a105-0c273ee0a428n%40googlegroups.com.

[AndresJ89]

Thanks a lot Laurent!