Network Optimization For An Urban Passenger-and-Package Sharing System

The prevalence of online retailing has led to the expansion of urban logistics systems and the growth of freight traffic,bringing significant negative impacts like road congestion and air pollution.As an efficient pathway to address such negative impacts,passenger-and-package sharing(PPS),which means passengers and packages are transported integrally by urban public transport systems,is receiving growing attention recently.Existing studies and practices mainly focus on single-modal PPS systems based on either transit lines or taxis,whereas they are facing issues like limited user participation,service coverage and resource utilization.Therefore,this study proposes a multimodal PPS system that utilizes the advantages of different transport modes.The proposed multimodal PPS system integrates metro and taxi modes with a huband-spoke network structure.Metro lines serve backbone links due to their high capacity,long transport distance and operation stability.Taxis serve spoke links due to their wide service coverage with flexibility.Hubs are established at metro stations to transfer packages between the two modes.Incentives are provided for taxi drivers and passengers to stimulate them to transport packages along their trips.The hub-and-spoke network structure is widely accepted in the logistics industry.This study focuses on the network planning problem for the urban PPS system,which concerns: 1)the re-equilibrium of the taxi network,and 2)the optimization of the hub-and-spoke PPS network.The first problem analyzes the influences of package transport services on the equilibrium state between passengers and taxis,which lead to a new equilibrium state among passengers,packages,and taxis.The second problem optimally determines the location of hubs and assignment of package flows on links,with the objective of maximizing the system profit.The main contents of this study include:(1)A framework for the urban PPS system.The structure of the interactions regarding payments,services and information among stakeholders in the PPS system is first proposed.Detailed interactions under collaborative and competitive business models are further analyzed and the collaborative model is adopted for this study.Then we introduce the operational process of the PPS system and the influences of the incentive strategy,which provide the basic scenario for modeling.(2)The PPS network optimization model with a stable taxi network.This part assumes that the incentive provided by the system improves the participation of passengers and taxi drivers in PPS systems,but does not lead to re-equilibrium of the taxi network.With a given capacity-incentive function,the model optimizes the location of hubs,assignment of package flows,and the incentives on spoke links.The model is a mixed integer non-convex programming problem.Due to the fact that the computational time is long and the global optimality is not guaranteed,we transform the model into a mixed integer convex programming problem and propose customized exact and approximate algorithms.The model and algorithms are validated using a standard dataset.The topological and economical features of the PPS network are demonstrated by comparing it with the traditional hub-and-spoke network.(3)The PPS network optimization model with taxi network re-equilibrium.In this part,we first propose a taxi network re-equilibrium model considering passengers,packages and taxi drivers,which captures the influences of package transport services and incentives on taxi drivers’ and passengers’ behaviors.Then,a master problem that optimizes the location of hubs and assignment of package flows is developed with the objective of maximizing the system profit.The taxi network re-equilibrium model is treated as a subproblem that outputs taxi capacities for the master problem.The subproblem and master problem are respectively solved using a hybrid fixed-point algorithm,which outputs exact solutions,and a successive linear approximation algorithm,which outputs approximate solutions.The latter utilizes the former for fast calculating parameters in the solution process.(4)Real-world case study of the PPS system in Shanghai.The applicability of the proposed model is demonstrated by comparing the model outputs with the real-world data.Then,the taxi-only and multimodal metro-and-taxi PPS systems are optimized and analyzed.Compared to a traditional system where passengers and packages are transported independently,the taxi-only PPS system improves the profit by 22.4%,reduces the average passenger travel fare by 4.3%,and improves the average driver income by 22.3%.The multimodal metro-and-taxi PPS network achieves better performance.It improves the profit by 61.2%,reduces the average passenger travel fare by 5.7%,and improves the average driver income by 34.0%.Through framework design,modeling and case study,this study proposes a set of network optimization methodologies for an urban multimodal PPS system integrating metro and taxi modes.The methodologies support the city manager in planning and analyzing the PPS network.