Ride-sourcing Vehicle Dispatching Strategy Considering Passengers' Confirmed-order Cancellaiton Behavior

With the rapid development of information technology,ride-sourcing apps,i.e.,Didi,Uber,and Lyft,quickly become popular in recent years.Massive real-time ride requests are dispatched every second on these ride-sourcing platforms today,but an unneglectable amount of confirmed orders(COs)are later cancelled by passengers.Passengers' CO cancellation behavior not only decreases the availability of drivers,but also lowers the dispatching efficiency and reduces revenue of the platform.In order to reduce the negative impacts of passengers' CO cancellation behavior,this thesis models passengers' CO cancellation probability due to the appearance of the vacant taxis,and proposes a driver-rider matching strategy considering passengers' CO cancellation probability.A simulation environment is built to examine the effectiveness of the proposed dispatching strategy,based on an open dataset of Chengdu's ride-sourcing market provided by Didi Chuxing.The simulation results show that the proposed dispatching strategy can effectively improve the platform's revenue when the ride-sourcing fare rate is high and the supply-demand ratio is small.However,when both the vehicle fare rate and the supply-demand ratio are low,the proposed dispatching strategy may reduce the dispatching revenue for being over-conservative,when compared to the dispatching strategy without considering passengers' CO cancellation behavior.And when the vehicle fare rate is low and the supply-demand ratio is high,those cancelled short-distance orders generate a higher profit than that of completed short-distance orders,so the proposed matching strategy would be incline to induce passengers to cancel those short-distance orders to earn the cancellation fee,which damages the interests of passengers and is bad for the long-term development of the platform.