Innovative Operations and Network Designs for High-Velocity Intra-City Courier Services

This dissertation introduces a novel network design for intra-city courier services, aiming to improve operational efficiency and service quality of courier companies. Traditional hierarchical network models in courier services mandate all shipments be aggregated at centralized distribution centers for sorting, which can create bottlenecks and delays. In contrast, this study proposes an alternative that utilizes the existing infrastructure of courier package drop-off and pick-up stores dispersed throughout urban areas. This alternative network design reorganizes these stores as mini sorting hubs, enabling the decentralization of the sorting process. The research is presented in three separate articles, each addressing different facets of the proposed network design under various conditions of demand, capacity, and operational constraints. In the first article, we examine tactical and operational planning for the proposed network structure. Our tactical approach uses a multi-commodity service network design to maximize consolidation and optimize commodity paths, managed through mixed integer programming. We refine these paths in a secondary model to minimize necessary cycles for service guarantees. In operational planning, we adjust our strategies based on short-term demand deviations, enhancing service levels through plans tailored to daily operational specifics. In the second article, we address the stochastic service network design for an intra-city courier service using a hybrid fleet of contracted and crowdsourced drivers. We strategically acquire capacity at reduced rates considering future demand and adjust dynamically based on actual crowdshipper capacities and spot market conditions. Our modeling approach uses two-stage stochastic programming with advanced decomposition methods, improving efficiency and adaptability to operational data. In the third article, we focus on service network design challenges for an intra-city express delivery system with specific hub capacity constraints. We model and solve the network design on a time-space framework as an integer program, using commercial solvers for smaller scenarios and a constructive metaheuristic approach for larger cases. This strategy separates the problem into freight routing and vehicle scheduling, iteratively solved to create robust solutions suitable for diverse real-world conditions.