| In recent years,as people’s vehicle travel demand continue to increase,some urban problems,such as traffic congestion,parking difficulties and environmental pollution,have become increasingly exacerbated.Electric Vehicle(EV)sharing has gradually become an emerging travel mode in major cities in China.However,at present,EV sharing is in their infancy,especially for one-way EV sharing which have been adopted by many companies.There are mainly problems such as the imbalance between supply and demand at each one-way EV sharing station and the high relocation cost in the operation.To solve these problems,this study proposes a vehicle relocation optimization method based on the users’ choice behavior.First,based on the survey and analysis of the operation and relocation mode of EV sharing,the mechanism of the users’ choice behavior is explored.Then,based on the users’ choice behavior,the vehicle relocation optimization under static demand and dynamic demand are studied separately,and some advice are provided for the operation of EV sharing companies.First,through the Stated Preference(SP)survey and questionnaire data statistics,the factors affecting the users’ choice behavior are analyzed.Based on the results of statistical analysis,a Nested Logit(NL)model is constructed to jointly analyze the mode choice and EV sharing station choice based on the characteristics of the four travel modes,namely public transportation,traditional taxis and online car-hailing,private cars and shared electric vehicles.Secondly,the vehicle relocation optimization method under static user demand is proposed.Under the assumption that the total number of EV sharing users throughout the day is determined,the full-day operation is divided into three parts according to the time distribution characteristics of users’ travel demand.In each time period,a dynamic discount strategy based on EV sharing station choice is proposed,which can guide reserved users to change the nearest default station choice behavior.A vehicle relocation optimization model is built at the end of each time period to minimize the total relocation cost of users and the personnel,and the tabu search algorithm is applied to solve the problem.Then,the vehicle relocation optimization method under dynamic user demand is proposed.Under the assumption that the total number of all-day travelers is determined,the service level of EV sharing is dynamically adjusted to attract some travelers who originally chose other modes to choose EV sharing,thereby obtaining the dynamic user demand based on mode transfer besides the original user demand.The full-day operation is divided into three parts according to the time distribution characteristics of users’ travel demand.In order to achieve the supply and demand balance at each station at the beginning of the next time period,a dynamic discount strategy based on the joint choice of travel mode and EV sharing station is proposed,which can guide reserved users and the dynamic users based on mode transfer to change the nearest default station choice behavior.A vehicle relocation optimization model is built at the end of each time period to minimize the total relocation cost,and the tabu search algorithm is applied to solve the problem.Finally,in the case study of some EV sharing stations in Haidian District,Beijing,the relocation optimization effectiveness is analyzed,that is,the total vehicle relocation cost,the personnel relocation cost and the personnel relocation tasks with and without dynamic discount strategy under static and dynamic user demand are compared and analyzed.The analysis shows that compared with non-dynamic discount strategy,the total relocation cost,personnel relocation cost and personnel relocation tasks with the corresponding dynamic discount strategy under static user demand are reduced by 4.82%,20.60%,and 10.71%,respectively.Compared with the non-dynamic discount strategy,the total relocation cost,personnel relocation cost and personnel relocation tasks with the corresponding dynamic discount strategy under dynamic user demand are reduced by5.83%,22.78%,and 14.58%,respectively.These results shows that the vehicle relocation optimization method proposed in this paper is reasonable and effective.Meanwhile,the impact of service level optimization ratio threshold and minimum dynamic discount on relocation cost is explored.The results show that only by setting appropriate service level optimization ratio threshold and minimum dynamic discount,and taking corresponding service level optimization measures can the total vehicle relocation cost be minimized. |
