| The stability of public transportation operations is a key indicator of service quality,and improving it is beneficial for achieving the policy of public transport priority and helping to alleviate urban traffic congestion.However,various factors such as weather,passengers boarding and alighting,and road traffic conditions often affect daily public transportation operations,leading to unstable phenomena such as long gaps and bus bunching,which seriously affect the efficiency of urban public transportation services.With the continuous development of information technology,public transportation trajectory data can reflect real-time information on bus operations,providing a guarantee for studying bus operation stability.Traditional research on bus operation stability mainly focuses on single routes,and there is still much room for improvement in the stability of the public transportation network.Based on this,this study uses continuous public transportation trajectory big data to further study the dynamic spatiotemporal characteristics of the public transportation network and deeply explore the mechanism of public transportation operation instability.This achieves the development of research on public transportation stability from a single route to a network and provides a theoretical basis and optimization suggestions for public transportation operation management.This study uses a large amount of GPS trajectory data from buses and employs data mining,machine learning,and regression analysis methods to delve into the dynamic spatiotemporal characteristics of the public transportation network.This study constructs a complete set of analysis indicators for the operating status of a single public transportation route and the operating status analysis indicator system of the public transportation network.It applies machine learning and geographic information systems technology to propose methods for predicting the travel time of a single public transportation route and the factors influencing the spatiotemporal characteristics of the public transportation network.By establishing reasonable prediction and regression models,this study provides a theoretical basis for subsequent empirical analysis and optimization strategies to improve the service quality and efficiency of urban public transportation,making a positive contribution to building smart cities and green low-carbon transportation.The main research results are as follows:(1)The bus travel time services fluctuate significantly during peak hours,particularly in the morning rush.Bus bunching and large headway incidents are prone to occur,and once they do,these phenomena persist until the destination is reached,making it difficult to resolve.In this study,a Kalman Filter-LSTM model is established for travel time prediction,which outperforms ensemble learning models and single LSTM models,achieving an R~2of 0.9.(2)The average dwell time in Jinan’s public transport system is relatively short,mainly concentrated between 15 and 35 seconds.However,the average headway at stops during peak hours exceeds the prescribed time by more than double,and the average headway at stops during peak hours is lower than during off-peak hours.The average headway at bus stops in Jinan’s central urban area is lower than in the surrounding peripheral areas.Moreover,analysis reveals that some bus lines experience wasteful departure intervals during off-peak hours,necessitating further optimization of bus schedules and vehicle dispatching to improve the operational efficiency and service quality of the public transport system.(3)The impact of the topological characteristics of the public transport network on the dynamic spatiotemporal features of the network operation is greater than the impact of the surrounding socio-geographic environmental features of the bus stops.This indicates that government departments responsible for transportation planning should place greater emphasis on the design and optimization of public transport network topology.Furthermore,when planning and constructing public transport networks in urban fringe areas,the distinct socio-geographic environments of these regions should be taken into account,adopting corresponding strategies and measures to ensure the sustainable development of the public transport network. |
