Food Order Delivery Using Crowdshipping: A MILP-based Rolling Horizon Approach

In recent years, the advancement in technologies has led to creative solutions like crowdshipping which deals with the use of occasional drivers in the delivery systems to lower the cost of transportation, and a small compensation fee is paid to the occasional drivers for their service. This paper considers a food delivery system, which incorporates occasional drivers in addition to dedicated vehicles present at each restaurant and makes decisions about the assignment of orders while minimizing the total cost and delivering the order within its time window. Here, each occasional driver has a detour radius and order delivery can be done by a single occasional driver or by transfer between two occasional drivers respecting their detour radii or by a restaurant’s dedicated vehicle. An event-based rolling horizon approach is proposed, at each decision epoch a mixed-integer linear programming model is used to solve the problem, and if the dedicated vehicles are used for the delivery, then a routing algorithm is implemented to build dedicated vehicles’ routes. The model guarantees the delivery of orders to the customers and a penalty is imposed if the delivery time exceeds the order’s time window. CPLEX solver is used to implement this model and computational analysis done on the instances demonstrates the impact of different configurations of restaurants, occasional drivers, and customers on key performance metrics such as total cost, order fulfillment rate, on-time arrival rate, and pending orders rate. The results show that this approach provides good solutions within short run times and transfer delivery by occasional drivers is a cost-effective way to reduce overall delivery cost.