| Trip-sharing is to match similar trips in time and space,and dispatch a vehicle to provide driving services for similar multiple trips.Taxi trip-sharing is of great significance to alleviate urban traffic congestion and improve the overall traffic efficiency of the city.With the popularity of the concept of trip-sharing,taxi trip-sharing has become the trend of future development.Different from carpooling(the trip’s demand is obtained in advance and the driving route is fixed),the trip’s demand of taxi passengers is generated dynamically.The driving route and status of taxis change over time.Dispatching a specific taxi to provide trip-sharing services for multiple trip requests required not only to meet the latest arrival time demand of passengers,but also to improve the time efficiency of the scheduling algorithm as much as possible and reduce the time of passengers waiting for the response of dispatching algorithm.When considering the real-time traffic condition of the road network,the amount of calculation of trip-sharing matching will become larger,and the planning of driving route will be more complex.Therefore,it is difficult to match large-scale trips and taxi dispatching at the city level.The main work of this paper is as followed:(1)In order to reduce the amount of calculation of matching relationship between taxi-trips,a fast screening strategy based on time,distance and direction constraints is proposed.The basic idea is to filter out most of the trips that cannot be shared with a given trip according to time and distance constraints,and obtained a small set of candidate trips.The strategy is applied to the trip-sharing matching method based on graph method,and the effectiveness is verified by using the taxi data sets of New York and Qingdao.The results showed that the filtering strategy proposed in this paper can greatly reduce the matching time without affecting the matching effect(the matching time of a single trip is reduced by about 82% and 92% respectively).Based on the scenario analysis of taxi trip-sharing in Qingdao,it is found that when the time window is set to 1min and the delay time is 5min,the theoretical upper limit of the proportion of the trips which joined in trip-sharing in working days is 45%,and the theoretical upper limit of trip-sharing total saving travel time and total saving travel distance that can reach about 5000 hours and 80000 kilometers,respectively.The results of weekend days and working days are very similar.(2)Taxi trip-sharing dispatching process is mainly divided into taxi searching and taxi matching.In order to reduce the computational complexity of taxi search,most research ignored the actual road conditions,and set the driving time of the road network as a static value.This paper proposed a method for estimating the road driving time based on historical trajectory,which is used to search the candidate taxis.When determining whether the candidate taxis can provide trip-sharing services for multiple trips,the driving route planning needed to be calculated in traversal operation,which will increase the amount of calculation.In this paper,the principle of "Monotonicity" is introduced,and a driving planning dispatching strategy that can terminate the traversal operation in advance is proposed to improve the calculation efficiency of taxi dispatching algorithm.On this basis,the taxi trip-sharing in Qingdao and Chengdu are analyzed.It is found that the average saving time,the proportion of shared trips and the saving total travel time in Chengdu are significantly higher than those in Qingdao on weekdays and weekend days.Through the proposed two indicators,namely,trip similarity and trip-sharing proximity,the empirical results are compared and analyzed. |
PDF Full Text Request