Collaborative matching of ridesplitting and connection in the ridesourcing market

The emergence of ridesplitting as a form of ridesourcing reduces the use of vehicles on the road.When connecting multiple ridesplitting orders and single orders to the sharing path,it can achieve higher sharing efficiency.This paper aims to further improve the vehicle sharing rate,and explore the impact of multi-mode sharing using the order data of Haikou,China provided by DiDi Chuxing.A shareability network combining the ridesplitting network and connection network is built based on the order data.We propose the on-demand matching algorithm with four matching objectives to obtain the dispatching strategy.The results show that the percentage of shared trips can reach 99%,the vehicle saving rate can reach 83%,and the average number of shared trips served by shared vehicles can reach about 6 with the time interval 20 min and maximum delay 300 s.When the maximum delay is 300 s,the percentage of orders that can be shared by multiple modes is about 30%.The average delay,idling time and waiting time of shared orders are slightly higher than the corresponding maximum delay,and increase with the increase of the maximum delay,while the change of saving time is the opposite.The proposed algorithm considers the impact of the maximum delay,which,compared with the maximum matching algorithm,has a significant improvement.

Quantifying the Impact of Transportation Network Companies on Urban Congestion in a Medium Sized City

In the recent years,TNCs(transportation network companies)and on-demand ridesharing services have grown rapidly.Given conflicting reports on TNC impacts,a need exists to study mode choice shifts in the presence of TNC services and their effects on urban congestion.Using Birmingham,AL(Alabama)as a case study,this paper showcases the feasibility of modeling TNC services using the MATSim(Multi-Agent Transport Simulation)platform,and evaluating the impact of such services on traffic operations.Data used for the study were gathered from Uber drivers and riders through surveys,as well as the US Census.The results indicate that when 200,400,and 800 TNC vehicles are added to the network,the VKT(vehicle kilometers traveled)increase by 22%,23.6%,and 23.2%,respectively,compared to the baseline scenario(no TNC service).Analysis of hourly average speeds,hourly average travel times,and hourly volumes along study corridors further indicate that TNC services increase traffic congestion,in particular,during the AM/PM peak periods.Moreover,the study shows that the optimal TNC fleet size for the Birmingham region is 400 to 500 active TNC vehicles per day.Such fleet size minimizes idle time and the number of TNC vehicles hovering,which have adverse impacts on TNC drivers,and the environment while ensuring TNC service availability and reasonable waiting times for TNC customers.