Research On Path Selection Model Based On Utility Regret Joint Rule

Transportation travel is an indispensable part of people’s participation in urban social and economic activities.The improvement of travel experience is significant to realize the refined management of urban transportation and improve the quality of the city.The realization of reasonable and accurate travel paths recommendation is one of the important steps to improve the travel experience.However,the current common travel service recommendation systems often ignore the users’ travel habits,and only provide the users with travel paths recommendation according to the uniform rules such as the shortest path or the shortest driving time,so it can not achieve reasonable travel paths recommendation.In recent years,with the deep integration of mobile Internet,Internet of things,GIS,big data and other technologies with urban intelligent transportation system,large-scale urban traffic trajectory data has been able to be efficiently collected,transmitted and stored.The urban traffic trajectory data implies the travel rule of urban residents and reflects the dynamic traffic state of urban road network.Based on the large-scale urban traffic trajectory data,it has become an urgent need in the field of intelligent transportation to analyze the route selection behavior of travelers,establish an effective route selection model,and use group intelligence to provide efficient and reasonable route recommendation for travelers.Under the cost-benefit constraints,urban taxi drivers usually have better path selection behavior than other residents,so it is a feasible solution to provide path recommendation for other travelers according to the path selection habits of urban taxi drivers.In this paper,the large-scale urban taxi trajectory datas are taken as the research object.Aiming at the problem that travelers can’t satisfy the complete rational hypothesis of utility maximization model when they choose paths,a path selection model under the utility regret joint rule is proposed by combining the regret minimization model under the limited rational condition with the utility model.The experimental results show that the probability of path selection calculated by the above model is closer to the probability of path selection in the real scenes than the utility model or the regret minimization model.Based on the above model,a path recommendation method based on the difference time utility regret joint model is proposed,which can recommend more suitable travel paths for the users.The specific research work of this paper is as follows:(1)Using the big data platform spark to clean the taxi trajectory data,map matching,trajectory segmentation,OD extraction and attribute value calculation of continuous passenger carrying trajectory segments,as the data basis of path selection model and path recommendation.(2)This paper proposed a path selection model based on the utility regret joint rule(utility regret joint model for short).This model combines the utility model of complete rationality with the regret model of incomplete rationality,and selects a group of utility regret joint models under different combination of multiple attributes according to the historical track data,and then determines the proportion coefficient.The experimental results show that the path selection probability of utility regret joint model is closer to that of real scenario than that of utility model and regret model.(3)This paper proposed a path recommendation method under the difference time utility regret joint rule.According to the four time segments of weekday / weekend,peak / off peak combination,the historical track data is divided and the utility regret model proportion coefficient under different time segments is modified.Finally,the path is recommended to users based on the historical track data of different time segments and the time difference utility regret joint model.The experimental results show that this method is more diversified than the current recommended path of navigation software and more consistent with the taxi habit path selection results.