Abstract

Loading and Unloading Zones (LUZs) are among the most pragmatic alternatives to reduce unregulated parking by freight vehicles. These designated zones provide temporary parking for freight vehicles to load and unload goods without obstructing traffic, occupying pedestrian areas, or causing double parking. Despite their relevance, existing research focuses on minimizing subordinate metrics such as walking distance or the number of LUZs, thereby neglecting the main direct objective of implementing LUZs in practice: reducing unregulated parking. In this paper, we aim to close this practical and methodological gap by proposing a novel Mixed-Integer Linear Programming (MILP) model for designing urban distribution networks with LUZs, simultaneously determining their location and size to minimize unregulated parking. To inform location and sizing decisions, our model also incorporates scheduling decisions for vehicle parking at LUZs, thus capturing both temporal and spatial capacity constraints. To enhance computational performance, we introduce valid inequalities that allow us to solve real-world, large-scale problem instances to optimality. To illustrate the applicability of our model, we conducted a real-world case study in Luxembourg City, involving fieldwork to characterize nearly 100 LUZs and 600 establishments. Extensive numerical experiments show that our model can solve these large-scale instances to optimality within minutes. From a practical standpoint, we derive several managerial insights into the design of LUZ networks, showing that the duration of the delivery horizon and the maximum walking distance to establishments significantly influence unregulated parking.