CVRP - > To put a capacity constraint for total demand of visited points located in last visited city

[burbot48]

Hi All,
I want to set a constraint for the vehicle capacity. The vehicle can visit multiple points existed in different cities. For each vehicle, the total demand of the points visited by the vehicle which are existing in that last city is expected to be more than half of the total demand within that vehicle. Can we put such a hard constraint.

I added data['cities'] variable as below to define which nodes are in same city.
data['cities'] = [0, 1, 1, 2] # Index 0 is depot, index 1and index 2 are in same city, index 3 is in a different city

I added a constraint like as below.

def last_city_demand_constraint(routing, manager, data):
        for vehicle_id in range(data['num_vehicles']):
            end_index = routing.End(vehicle_id)
            capacity_dimension = routing.GetDimensionOrDie('Capacity')
            total_demand_in_vehicle=capacity_dimension.CumulVar(manager.NodeToIndex(end_index ))
            demand_vars = []
            for node in range(len(data['cities'])):
                if data['cities'][node] == data['cities'][manager.IndexToNode(end_index)]:
                    demand_var = capacity_dimension.CumulVar(manager.NodeToIndex(node))
                    demand_vars.append(demand_var)
            last_city_demand = routing.solver().Sum(demand_vars)
            routing.solver().Add(last_city_demand >= total_demand_in_vehicle/2)
 
    last_city_demand_constraint(routing, manager, data)

But the solver status is coming Infeasible. Can anybody check why that constraint is not working properly.

Thank you

[burbot48]

Is there anybody had an idea to manage that constraint.
Thank you