Real-time Control Strategy For Urban Public Transportation Oriented Different Operation Environment

An efficient,convenient and comfortable public transportation system is an effective way to solve urban transportation problems.With the improvement of the level of urban motorization and the increase of the density of the transit network,optimizing the transit network and increasing the fleet resource supply from the planning level have been unable to significantly improve the service level of the public transportation system.Because buses are inevitably interfered by a variety of internal and external factors in actual operations,such as road traffic conditions,unbalanced transit demand in time and space,traffic accidents.Buses cannot operate completely according to the predesigned schedule and exert their approved transportation capacity.Therefore,it is particularly important for the high-quality development of public transportation systems to maximize the service efficiency of transit resources and improve the service level of urban transit through real-time control methods.For cities with a well-developed public transportation system,bus lines are dense and the spatial layout of the lines is diverse,which makes the lines in different operating environments.Both internal and external operating environments will impact on the daily operating status of buses and the design and implementation of real-time control decisions.Through in-depth analysis of the operating environment characteristics of the bus lines,this paper divides the bus lines into three types:a single bus line,multiple common bus lines and multiple bus lines oriented to the bus hub.Real-time control strategies adapted to the line operation environment are designed respectively for each type of bus line.Aiming at maximizing the benefits of stakeholders,real-time control models for buses are constructed and the corresponding solution algorithms are proposed.This study hopes to enrich the theoretical framework of real-time bus control problems.The main researches of this paper are as follows:(1)A real-time hybrid control strategy for buses on a single bus line.The single bus line in this study refers to a bus line whose operating environment is relatively independent.With the analysis of the characteristics of the real-time control problem of the single-line,and the fully consideration of the uncertainties of various influencing factors during the bus operations,this study proposed a data-driven hybrid control framework that combines the rich experience of dispatchers with theoretical optimization models.The decision-making behavior of dispatchers are extracted and abstracted,and a machine learning method is used to simulate the decision-making process.Considering the capacity and service quality of buses,a mathematical planning model is constructed with the goal of minimizing passenger travel costs.A real-time control mechanism including three stages,control timing judgment,control strategy selection and control action parameter optimization is formed.Further,in order to explore the impact of future control decisions on current decisions,two methods based on local and global are used to solve the model.(2)The real-time bus control strategy under common bus lines environment.In this study,the common lines are defined as two or more lines that share part of the bus corridors,and the overlapping part needs to account for 50%or more of the entire line.In-depth analysis of the demand and operating characteristics of the common lines,considering the interaction mechanism between the passenger riding plan choice behavior and the bus control decisions,and combining with the relevant research results of behavior theory,the passenger riding plan choice behavior is characterized.The passenger travel plan choice behavior model is embeded into the dynamic bus dispatching time control model,to establish a feedback mechanism including bus control,real-time information release and travel plan decision.The influence of real-time information release,travel plan selection,and operation mode on the implementation effect of control dicisions is discussed.From the perspective of competition and cooperation,with the goal of maximizing the generalized benefits of bus operation participants,a real-time control model considering passenger travel plan choice for common lines is constructed.Finilly,an algorithun is designed and the actual bus line is taken as an example for empirical analysis.(3)A coordination control strategy for buses on multiple intersecting bus lines.For a group of lines that intersect at a bus hub and belong to the same bus operator,this study hopes to make full use of the characteristics of their operational structure—the structural connectivity between the lines,and the heterogeneity of each line's operation in space and time,to design a coordination control strategy.The cross-line operation method realizes the replacement of buses at the hub station,and uses the concept of "sharing economy" to improve the utilization of fleet resources and the overall service level of the multi-line transit system."Coordination" has two meanings in this study:one is to coordinate the fleet resources of multiple lines;the second is to coordinate the holding control,short turning,and cross-line control strategies.With the full consideration of the positive and negative impacts that real-time control strategies may bring to passengers,this study proposed the coordination control model for multiple bus lines intersecting at a bus hub.Then,a column generation algorithm is designed to solve the model.The effectiveness of the model and algorithm are verified according to the actual operating data,thereby investigating the ability of the multi-line coordination control strategy to resist large and small random interference.