Bus Operation Characteristics Analysis And Management Control Strategies Design Based On Microscopic Simulation

With high occupancy and low per capita energy consumption,public transit is one of the most important transportation modes for residents.However,most of the bus route systems at present suffer from many problems,such as slowly moving,seriously crowded environment in buses and at stops,fluctuating travel times.How to improve the operation efficiency and service level is the focus of the issue for urban planner and public transport manager.The key to solving the problem is to reduce the unnecessary interference and delay during the bus operation,for example buses clustering round buses(i.e.bus bunching)phenomenon,and buses’ needless dwelling at stations.To achieve this goal,this thesis understands the characteristics of the passenger demand distribution and the interference mechanisms of social vehicles,and further designs the optimal bus control strategy.Besides,it is greatly necessary to find out the factors which affect the behavior of the morning peak-period commuters and how the factors act,and then design and evaluate the transport policy.This paper applies both simulation method and optimization method;and utilizes the expertise of traffic engineering,operational research,computer science and other disciplines.These researches on the modelling of bus running,mechanisms of bus operation and relative phenomena,design and optimization of bus control strategies,and passenger travel behaviors are carried out.The achievements of our works can ultimately provide the theoretical foundation for the transit planning,design and operation management.The researches of this manuscript mainly includes the following aspects:(1)The bus running data are extracted and matched,and the bus route model is investigated and verified,the bus operation characteristics are also analyzed.The study associates the bus GPS data,the passenger IC card data,and bus stop data,and obtains the bus operation related information including the bus stopping time,passenger boarding time and passenger alighting time.Based on the information,a logit-based algorithm for the estimation of bus dwelling times at stops is designed and verified.The researches show that the investigation based on the multi-resource data contributes to the reflection of the bus running states.Besides,the proposed fit method is theoretical-guiding significance to depict the bus running in the real-world.(2)A bus operating optimization model considering A/B skip-stop strategy applying are constructed.This study focuses on a two-way bus line operating on a dedicated bus lane,and proposes a skip-stop strategy making an all-around consideration of the passenger demand distribution on both directions.To simplify the implementation,we adopt the A/B skip-stop service.This service defines three types of stops:A,B and AB.In the service,the buses depart alternately from the original stop as type A and B,and A(or B)buses serve A(or B)stops,as well as AB stops.Besides,each passenger is assigned to a fixed origin and destination before his(or her)trip,and he(or she)will go directly or take transfer to destination stop under the skip-stop service.Taken together,a simulation-based optimization method to design A/B skip-stop service is proposed,where we have considered each passenger’s origin and destination.A genetic algorithm is then developed to seek for a best service pattern when the passengers’ travel time is minimized.The example of real-world bus line indicates that the bidirectional skip-stop strategy outperforms the unidirectional one in terms of average travel time saved for passengers.Simulation results also suggested that the bidirectional skip-stop strategy can reduce bus bunching,improve the in-vehicle congestion and increase the running stability of the buses.Finally,the elastic demand that transferring passengers may change their origins or destinations has been considered,it is found that the passengers’ travel choices impact the optimal stopping schemes.(3)The bus route model built on a two-lane mixed traffic road is developed and evaluated.The fact that the bus lanes are mixed with social vehicles in cities are taken into consideration,and a cellular automaton based model is proposed to depict the interaction between cars and buses.In the model,social cars can change their lanes on both lanes,while buses are only allowed to run on one lane.To reduce the bus bunches,a threshold-based integrated strategy of holding and limited boarding control is proposed.In the bus dwelling time,if the time gap between the current bus and the front bus is too large or too smaller,a limited boarding control or holding control will be triggered.The model parameters are calibrated with the data collected from a real-life bus route,and the simulation results show that the optimal control settings vary with the congestion level.Besides,a good bus control strategy not only improves the efficiencies of bus operating and passenger travel,but also speeds up the car running by alleviating traffic congestion.It is also worth noting that,traffic participants,such as passengers,buses and cars,can enjoy larger travel time savings or speed increases by the integrated strategy under more crowded conditions.(4)The morning peak-period commuting behaviors under skip-stop service are investigated.The transit commuters’ costs under the A/B skip stop service and the no-control scenario are compared,and a regret-minimum based Q-leaming algorithm considering global information is developed to depict the morning commuters’reinforcement learning characteristics.Firstly,the effectiveness of the algorithm is validated on an ideal many-to-one bus route.Then the effects of global information are further investigated,and the commuters’ departure time selection and corresponding commuting reward under different scenarios(e.g.,different locations of skipped stops,different bus capacities,and different numbers of bus stops)are analyzed.The research fills the blank of theoretical analysis about the effects of control strategies on the passenger travel behavior characteristics.