Optimization Of Taxi Sharing Path Based On Trajectory Big Data

The urban transportation system affects the rhythm of urban development.With the development of the urban economy and the continuous changes in human living habits,people’s travel demand has increased year by year,and the pressure on the urban passenger transportation system has continued to increase.As an important supplement and composition of urban passenger transportation,taxis also undertake more arduous transportation tasks.Taxi ride-sharing is an effective way to improve the efficiency of taxi transportation and alleviate the pressure of urban traffic.It is also a feasible way to increase the capacity of the urban passenger transportation system without basically changing the existing traffic.Carpooling is also important for reducing traffic pollution and reducing energy consumption.This paper first preprocesses the trajectory data of massive taxis in Shanghai,and obtains the OD data of taxis.Spatial clustering of OD data of taxis by clustering algorithm,dividing areas,and statistical demand;using artificial neural network model to predict demand data;establishing taxi ridesharing model and route selection algorithm considering future predicted demand and verifying the model and algorithm The effectiveness.The main research contents are as follows:(1)Clean the massive taxi trajectory data in Shanghai,and deal with the lack,duplication,and information errors in the trajectory data.After analyzing and processing the trajectory data,the OD of the taxi is obtained.After that,the K-means clustering algorithm is used to spatially cluster the OD data of the taxi,and then the Tyson polygon generation algorithm is used to divide the area of Shanghai,and the OD data of each area is counted.(2)Divide the OD data into time series,and collect weather,holidays and other factors that obviously have an impact on travel for feature engineering,and use deep neural networks to predict regional demand data.(3)Aiming at the total cost of the system,design a taxi ride-sharing model and route selection algorithm that take into account future forecasted demand and potential income.Finally,based on Shanghai’s road network data and forecast demand data as tests to verify the effectiveness of the models and algorithms.This paper focuses on the problem of taxi ride-sharing routes and considers the potential future demand.In order to better predict demand,taxi trajectory big data and neural network models are used.Apply forecasted demand to the path decision process.The experimental results show that the prediction results of big data can also effectively optimize the taxi ride path.