Bus Bridging Optimization For Urban Rail Transit Under Service Disruption With Passenger's Travel Behavior Analysis

Urban rail transit(URT)has become the dominant mode of urban development.However,as the scale of URT network continues to increase,the operating environment has become more and more complex.Once service disruption occurs,how to accurately grasp the travel behavior of interrupted passengers and formulate efficient and highquality bus bridging services to ensure their subsequent travel has become the focus issue of urban transportation management departments and URT operation managers in recent years.The traditional bus bridging service is mainly based on the passive relief connection after plenty of passengers have been stranded,and lacks consideration of the behavior of interrupted passengers and changes in elasticity of demand.Therefore,it is necessary to develop a travel plan choice model under URT service disruption considering the time and space characteristics,and then conduct the optimization of bus bridging on this basis to reduce the bus bridging operation and ensure interrupted passengers reach their destination station.The research of travel behavior analysis under URT service disruption and bus bridging optimization are summarized,and the pros and cons of the research are analyzed.In light of that,the behavior data is obtained by using the stated and revealed preference survey.The models of travel plan choice behavior considering uncertainty and heterogeneity are developed based on the obtained data,and the decision-making mechanism and influencing factors are analyzed.Then,taking into account the interests of passengers and operating departments,the optimization models of bus bridging service operation planning and scheduling are established based on elastic demand and joint dispatch.In detail,the research work and findings carried out in this paper are summarized as follows:(1)Investigation and characteristic analysis of travel plan choice behavior under URT service disruption.To be closer to realistic scenarios,a time interval,instead of a fixed value in previous studies,is set in each scenario of the stated preference questionnaire to indicate the uncertain duration of URT service disruption.Subsequently,based on the data obtained by an online questionnaire survey,the characteristics analysis of travel plan choice behavior under service disruption is conducted.And the result shows that the scene attributes(time interval of service disruption duration,occurrence period,passenger location),socio-demographic attributes(gender,income,age)and travel characteristics attributes(travel frequency,service disruption experiences,information service usage)may affect the travel plan choice behavior under URT service disruption.(2)Modeling travel plan choice behavior under service disruption with bus bridging considering uncertainty and heterogeneity.The attribute-specific extended expected utility(ASEEU)and risk measurement with Shannon entropy(RMSE)are used in the travel plan choice modeling to handle the uncertainty in the passenger travel plan caused by the uncertain service disruption duration.Furthermore,to analyze the heterogeneity of decision-making mechanism and perception weights,based on the latent class(LC)model framework,the LCASEEU&RMSE and LC-RMSE models are constructed.The results show that there are two classes of decision-making mechanisms: ASEEU and RMSE,and the proportion of the latter is much higher than the former.Meanwhile,passengers can be divided into two categories: risk-optimistic and risk-pessimistic,and risk-pessimistic people are more sensitive to the perception of waiting risk under service disruption.passengers who are younger than 35 years old,travel more than 10 times per week,have experienced service disruption,and use information service are more likely to be risk-optimistic;while women,passengers with a monthly income of more than 10,000,and passengers who arrive at the airport in advance are more likely to be risk-pessimistic.(3)Optimization of bus bridging operation planning with elastic demand.Based on the behavior model that can accurately describe the decision-making process of interrupted passengers,objective and high-precision predictions of passenger demand distribution under service disruption can be achieved.Meanwhile,taking into account the diversity of stops,a set of alternatives for the combined skip-stop bus is constructed.In light of these,the optimization of bus bridging operation planning with elastic demand is constructed.The discrete event simulation including the passengers and the bus bridging is designed and the simulation process is embedded in the genetic algorithm based on the mutual feedback process between the elastic demand and bus bridging operation plan considering the capacity limitation.At the same time,a doublelayer chromosome that can characterize the departure interval of buses with different stop plans is designed to improve the efficiency and accuracy of the solution.Combined with the historical service disruption event of Guangzhou URT,an experimental example is designed to verify the effectiveness of the constructed optimization model and the solution algorithm.It shows that compared with no bus bridging service(all-stop bus bridging service),the optimization results obtained by the proposed model in this paper can reduce 27.3% and 28.8%(9.6% and 11.7%)of the comprehensive cost respectively in the upstream and downstream directions.Sensitivity analysis is also conducted to provide policy recommendations for the URT management department.The results showed that: travel delay cost,operation cost,and comprehensive cost all increase with the increase of the overestimation coefficient of duration;the overall cost reveals a downward trend and the time-consuming an exponential increase with the increase in the number of stops plans;the overall cost rises and the rate increase as the bus bridging service is delayed.(4)Optimization of bus bridging service joint scheduling.The optimization framework for bus bridging service scheduling with the joint dispatch of buses with different types of stops is determined based on the analysis of the characteristics of bus bridging operation planning.Taking the dispatching empty time and the number of vehicles as the optimization goals,and considering the relevant time and space constraints and dispatching constraints,the optimization model of bus bridging service joint scheduling is constructed.The crossover and mutation operations in the NSGA-? algorithm are adaptively modified considering the multi-objective and wide solution space characteristics of the optimization to ensure the timeliness of the solution.Taking Guangzhou URT as an experimental case to verify the effect of the optimization model and the solution algorithm,the results of repeated experiments show that the solution algorithm has good timeliness and stability.At the same time,the optimization result under independent scheduling is set as an experimental comparison for joint scheduling.The results show that the Pareto optimal solution set of the proposed model with joint scheduling is better than that of independent scheduling.Meanwhile,without inserting empty trips,a joint scheduling solution in the Pareto set can reduce the use of 3vehicles compared with independent scheduling,which confirms the superiority of the joint scheduling strategy for bus bridging service.