Optimising Last-Mile Delivery with Mobile Automated Parcel Lockers Using Rolling Horizon Strategies

Parcel delivery in metropolitan areas using dynamically dispatched smart mobile automated parcel lockers (MAPLs) presents a logistical model for e-commerce that improves on traditional door-to-door and fixed-point pick-up methods. The study considers a problem where a retailer needs to decide daily on which customers to serve, how many MAPLs to dispatch, and where to position them without visibility into future orders. Failing to serve a customer with MAPLs within the maximum delivery lead-time necessitates the use of an undesired alternative, which is third-party delivery services. The challenge lies in minimising delivery costs while ensuring demand fulfilment under the dynamic daily arrival of orders. To address this sequential decision-making problem, three myopic mixed-integer linear programming models have been formulated within a multi-period rolling horizon framework. All models are evaluated and compared with each other. The findings highlight that the third strategy demonstrated a good tradeoff between delivery cost and delivery lead-time.