Multi-criteria decision making in dynamic slotting for attended home deliveries

Abstract To plan attended home deliveries efficiently and profitably, dynamic slotting can tailor the set of time slots offered to the customer. While most existing approaches focus on short-term criteria like revenue, marketing research highlights additional business objectives. For instance, distributing deliveries evenly across the region can ensure the visibility of branded trucks. Alternatively, prioritising influential customers can entice them to recommend the service to others. To accommodate additional objectives, we propose a multi-criteria approach that allocates delivery capacity to time slots and areas in two steps: First, considering spatial information on potential future customers and their values, it computes a predictive routing; second, it solves a revenue management problem to identify the most relevant customers given the set of slotting criteria. A comprehensive computational study demonstrates the effects for three exemplary criteria: revenue, visibility, and social influence. The results show that, depending on the demand scenario, additional criteria can be accommodated at little cost to revenue. However, the interaction of routing effects and offer set optimisation can cause unexpected shifts when changing the weight of criteria in the objective function. Hence, we propose a simulation-based process to let decision makers evaluate the effects of multi-criteria constellations.