On-line Dispatching Model Of Car-hailing With Incomplete Demand Information And Its Applications

As a new type of "Internet +" model,online car-hailing has been greatly developed,which has changed the traditional offline taxi service mode.The increasing number of users and the number of passengers The demand for convenient travel puts forward higher requirements on the dispatch strategy of online car-hailing platforms.The online carhailing dispatch problem refers to the complete process of the online car-hailing platform first solving the online driver and passenger matching problem based on demand information and online driver information,and then allowing the fulfilling vehicle to drive to the passenger’s destination in the shortest route.In the demand sequence,the interweaving of information such as the time of demand issuance and the passenger’s most evening bus time is highly uncertain,which makes it difficult to accurately predict the future demand distribution.For solving the problem of uncertain demand information,the usual thinking is to directly seek the optimal solution of the problem in the worst case,or to seek the optimal solution in the average sense after assuming that the demand obeys a certain distribution or random distribution.These two methods transform changing factors into offline problems to deal with.Some special cases of changing uncertain factors may give solutions that are far from the actual optimal solution.The theories and methods of online problem and competitive strategy do not make any assumptions about the future demand information,which can avoid the disadvantages of traditional offline methods that excessively rely on basic assumptions.Based on this,this article uses the theories and methods of online problems and competitive strategies to study the online car-hailing dispatching problem with unknown demand.First of all,for the online matching of the driver and the passenger.After the passenger sends a ride request through the APP,the platform will comprehensively consider the passenger demand information and the online driver’s information to determine whether the platform’s service capabilities can serve the demand in a timely manner,and ensure that the vehicle matching the demand must be at the passenger’s most evening time Before arriving at the passenger destination.Based on this,this article analyzes the online driver and passenger matching problem with service selection and service time constraints that completes the demand sequence in the shortest time,and divides the problem into platforms that can serve all needs in time,a demand that cannot be served in time and There are three situations where multiple demands cannot be served in time,and the online driver and passenger matching strategy and its execution effect are analyzed separately in these three cases,and the execution effect of the online driver and passenger matching strategy is analyzed through an example.Then,for the performance of the driver and the passenger after the match.In this thesis,the evolutionary game theory is used to construct a tripartite evolutionary game model of the online car-hailing platform,the driver and the passenger,and the reward and punishment mechanism of the online car-hailing platform is used to supervise the performance of both the driver and the passenger.Analyze the stability of the evolutionary game path of the three game subjects and the stability of the evolutionary strategy under the joint action of the three subjects,and then conduct a simulation analysis through Matlab.The results show that: only when the platform rewards both drivers and passengers only when the intensity and punishment are strong can both the driver and the passenger choose the performance strategy.Finally,for the emergency route selection problem of the fulfilling order-receiving vehicle on the way of picking up and driving off.This thesis analyzes the competitive ratio of emergency strategies such as waiting strategy,greedy strategy,and selection strategy when the order-receiving vehicle encounters sudden congestion on the way of picking up and driving off,and analyzes the effectiveness of these three strategies.The calculation example verifies the effectiveness of the above three strategies and their execution effect,and the results show that the execution effect of the selected strategy is the best.This thesis uses the theories and methods of online problems and competitive strategies to study the online car-hailing dispatching problem with unknown demand sequence.It not only provides a reference for the platform dispatching strategy,but also expands the application range of online algorithms and theoretical research on online algorithms.Providing theoretical support for the operation and management of the online car-hailing platform will also help increase the profit of the online car-hailing platform,thereby enhancing the industry’s competitiveness of the enterprise.