Customer Behaviour Analysis for Coordinated Delivery Service Selection Using Mobile Parcel Lockers and Autonomous Robots: A Kernelised-Support-Vector-Machine-Based Congestion-Game Model

Last-mile logistics operators have recently introduced mobile parcel lockers (MPLs) and autonomous delivery robots (ADRs) to alleviate traffic congestion and operational costs. Their ability to relocate their position during the day has the potential to improve customer accessibility and convenience, allowing customers to collect parcels at their preferred time among one of the multiple locations. Previous research on MPLs and ADRs primarily focuses on operations analysis and microscopic optimisation models (e.g., location-routing planning and task scheduling). A discussion of customer decision behaviour (e.g., service preferences and rationality levels) regarding MPLs and ADRs service selection and their effect on system efficiency has yet to be performed in the context of global information applied to customer decision patterns. Accordingly, this paper proposes a congestion-game model to investigate the performance of delivery systems that utilise MPLs and ADRs while taking into account a variety of operating parameters of the two delivery options (e.g., price, delivery efficiency). Additionally, we discuss how varying levels of customer rationality impact the performance of the system. By exploiting the global information of the delivery system (e.g., average cost estimation), we developed a novel algorithm that improves customer decision-making in service selections and system efficiency. The results show that customers are more sensitive to higher-cost and capacity-limited services (e.g., ADRs). The self-interested behaviour of customers will compromise the efficiency of the system when the equilibrium is achieved. Finally, the improved algorithm with global information applied increases the system performance by 28% to 68% (depending on customer rationality).