(Georgia Institute of Technology, 2021-06)
Bidoni, Zeynab Bahrami; Montreuil, Benoit
The rapid growth of demand and fierce competition are encouraging logistics service providers towards expanding their competency in terms of offering novel and faster services and reinventing their logistics system so as to profitably and sustainably gain market shares. However, analyzing customer behavior and the underlying causes of demand variability for new services are complex tasks. This paper is dealing with customer behavior modeling for a service provider who wants to extend its offering system to much faster delivery service than ever done before. To adjust its logistic capacities with future demand, it needs to estimate the volume and geographical distribution of demand for new offered services. By capturing customers’ sensitivities to the delivery-time observed in historical sales data and geo-categorization of orders in different time factors, a scenario-based demand generation methodology and tool are introduced for generating a wide range of demand scenarios with probabilistic patterns for customer behavior over all service offers with dynamic pricing. These are used to feed a simulator which models large-scale urban logistics networks service and offerings. In an application, it enables testing service capability improvements achievable by leveraging Physical Internet aligned transformation in a Chinese megacity.