Research On Task Assignment Model Of Ride-hailing Based On Mobile Scenario

With the innovative development of global positioning systems,5G,and smart devices,ride-hailing has become a mainstream trend in terms of travel.Today,passengers often have multiple modes of transportation from their place of departure to their destination.When choosing between these different modes of transport,passengers consider many criteria such as cost,travel time,flexibility(ability to adapt to schedule changes),convenience(such as location of starting point),reliability and safety sexual perception.Consider first systems such as buses or subways,which provide travel options with fixed geographic routes and fixed schedules.These fixed geographic routes charge a small fee to passengers,but bring little convenience.In contrast,private car or taxi services are more expensive but offer more flexible,more convenient and often faster options.As a new mode of travel,ride-hailing has changed the disadvantages of traditional offline taxis.Task assignment is one of the important contents of research on ridehailing.In the task assignment process,the combination of supply and demand forecasting,path planning,service points(driver service evaluation system),cloud computing,and “global optimization” of machine learning is used to realize order assignment.Efficient and convenient to meet people's travel needs has very important practical significance for people's lives.On the basis of reviewing the domestic and international research on ride-hailing and task assignment literature,it is found that there are few domestic researches on ride-hailing task assignment.About ride-hailing task assignment is one of the current research hotspots in the field of task assignment,Based on the analysis of the current situation of ride-hailing task assignment,there are mainly problems of space-time matching,imbalance between supply and demand,uncertain transportation capacity and load imbalance.Therefore,this paper mainly studies the problems of space-time matching and load imbalance,and constructs the ride-hailing task assignment models in the mobile scenario.Firstly,aiming at the problem of time-space matching between drivers and passengers,a task assignment model is constructed to meet the needs of ride-hailing platform.Compared with the random selection method and QSTA method,the greedy ant colony algorithm is more effective in the performance of the algorithm,the incentive cost,the optimization goal of task execution efficiency and the running time of the selection method.Then,this paper studies the driver load imbalance in the process of ride-hailing task assignment,and constructs a load balancing task assignment model to meet the ride-hailing platform.The demand of ride-hailing users includes the demand of participants,mobile users and ride-hailing platform.Therefore,the optimization objective function is the lower perceived cost of ride-hailing platform,the greater driver load balance and the shorter waiting time of mobile users.The model is solved by the discrete particle swarm algorithm of double scale mutation operator.Compared with the basic particle swarm optimization and genetic algorithm,the discrete particle swarm optimization with two-scale mutation operator has better convergence and optimization results.Finally,based on the problems existing in the task assignment of ride-hailing,and considering from the perspectives of government,ride-hailing platform,driver and passenger,the paper puts forward countermeasures and suggestions to promote the development of ridehailing platform,including realizing the time-space matching between passengers and drivers through refined operation and supply-demand prediction,increasing the number of ride-hailing and dynamic premium to achieve the balance between supply and demand of ride-hailing,and implementing employment driver system and optimization of capacity structure to solve the problem of ride-hailing platform capacity;optimization of scoring system and path recommendation to improve driver task load balance.