A template-based adaptive large neighborhood search for the consistent vehicle routing problem

The consistent vehicle routing problem (ConVRP), as defined in [1], is an extension of the periodic vehicle routing problem. This problem category involves the construction of routes over a given time period, e.g several days, such that customer demands are met. However, a simple delivery is nowadays not enough for companies to differentiate themselves in a competitive environment. For this reason, many companies started to focus on customer satisfaction to increase customer loyalty and to obtain a competitive advantage. From the customers’ point of view, an important criterion for a high quality service is consistency. So, the ConVRP, besides considering the traditional constraints, also accounts for this additional requirement of service consistency. This is modeled in two ways. First, in order to form a stronger relationship between the supplier and the customer, a customer can only be assigned to one driver. Second, to enable the customers to prepare themselves for a delivery, they have to be serviced at about the same time of the day. Customers to be serviced are situated at regionally different locations where they demand a predetermined number of services during the planning horizon. The demand pattern for each customer is characterized by the amount of goods delivered on a given day and the time needed for this service. A delivery can be required on several days, however, customers are visited only once a day if service is demanded. Visits are performed by a homogeneous fleet of vehicles. This fleet is located at a single depot, which is left at time 0. The number of vehicles is not restricted but their maximum capacity is. Due to the tour length restriction, the vehicles must return to the depot before the maximum tour length is exceeded. Even though the fleet is homogeneous, the drivers are not. The first consistency requirement therefore dictates that a customer is serviced by the same driver over the whole planning horizon. The second consistency requirement bounds the difference between the earliest and the latest arrival time over multiple days for all customers to a given value. Vehicle idling to decrease the arrival time difference is not allowed. The objective is to create a routing plan for each day in the planning horizon such that the aggregated travel time is minimal and the constraints are satisfied.