Urban Rail Service Design For Passenger And Freight Transport On Airport Lines

With growing traffic demand and the development of airports,many cities have implemented or planned urban rail lines to serve their local airports because they could provide green,efficient,and on-schedule service.Airport rail lines can be categorized into two types: airport-focused(AF)and non-airport-focused(NAF)rail lines.For NAF rail lines,this thesis proposes an integrated optimization method of train stopping and scheduling with express/local mode to satisfy commuting demand and provide passengers travelling to or from the airport with efficient train service.For AF rail lines,in order to enhance the operation profits of the rail lines,this thesis advocates to introduce freight shipments and proposes an optimization approach for freight train service design with mixed passenger and freight transport.The main contents and conclusions of the dissertation are shown as follows.(1)A comparison of line length,station density,passenger demand,and operational plans between the NAF and the AF rail lines is made.The results show that it is advisable to adopt express/local mode on the NAF rail line considering the demand from commuters and passengers to or from airports.The result also demonstrates that using AF rail lines to transport certain types of goods could be an alternative.A framework to evaluate the feasibility of freight on urban rail is proposed,which includes quantitative indicators and critical techniques needed throughout the process.Available transport formats of freight on airport lines are also investigated.(2)For NAF rail lines with express/local mode,express trains serve a certain number of stations while local trains stop at each station.Considering that passengers face a choice between with two types of train service,an integrated optimization model of train stopping and scheduling is designed with the objective of minimizing total passenger travel time.The decision variables consist of several binary variables which pertain to the stop schemes of express trains and overtaking stations,and several continuous variables,which represent dwell times at stations.The non-linear model has been transformed into a linear format through the use of linearization techniques,which enables it to be solved by a commercial linear solver.A guided branch-and-cut algorithm is designed by setting a special order of branches to accelerate the solving process.A numerical and a practical case study have been implemented to demonstrate the efficiency of the proposed method.Compared to the standard stopping pattern,passenger travel time decreases by 11.31% with express/local mode.A comparison is made between the optimized results produced by existing research and this paper based on the practical case study.As this thesis optimizes train stopping and scheduling simultaneously,the location of overtaking stations and timetables are accurately illustrated,which leads to a decrease in total passenger travel time.The performance of the guided branch-and-cut algorithm is also compared to the direct use of the commercial linear solver.The guided branch-and-cut algorithm obtains the best solution within 10 hours,which costs less time and produces a better solution than using the linear solver directly.(3)For AF rail lines with freight service,a strategy to advance operation profits is proposed by introducing freight service.Based on the existing passenger train timetables,an optimization model of freight train service design to maximize freight profit is constructed.Two types of freight transport are considered,which include both inserting freight trains into the operating plan and utilizing the extra space of passenger trains to serve freight demand.The decision variables of the proposed model refer to the number of train carriages,the train stopping schemes,and the train schedules.By introducing auxiliary variables,the non-linear model is transformed into a linear one.A numerical case study of the UP Express in Toronto is implemented to examine the efficiency of the proposed approach,and a comparison of optimized plans is also made between the existing researches and this work.The results show that the freight profit is $1,210 and 57.6% of freight shipments are served by using the freight train and two passenger trains to transport goods.The method presented in this Chapter,which considers two types of freight transport formats,can always obtain more efficient train service than existing approaches,which only consider one type of freight transport format.It is necessary to consider two types of freight transport formats to utilize the capacity of both passenger and freight trains.(4)For AF rail lines with freight service,when passenger train timetables are allowed to be slightly modified,a collaborative optimization model for passenger and freight train service is designed.This model is constructed to determine the headways and dwell times of passenger trains,with the objective to maximize profit,and is used to determine the number of train carriages,stop schemes and schedules of freight trains.Two heuristic algorithms are designed to enhance the applicability of the optimization method to largescale cases.Numerical case studies based on the UP Express and the New Airport Line in Beijing are conducted to demonstrate the efficiency of the proposed method.For the case of UP Express,compared to the optimization method with fixed passenger train schedules,the freight profit increases from $1,210 to $2,450 dollars,which finish all freight shipments in the research period.A sensitivity analysis of freight profit is also conducted.Increasing the price or decreasing the train operation cost could produce more profits,with profits being more sensitive to price than train operation costs.It is feasible to increase train capacity and loading efficiency,or relax the headway constraints,to further increase profits.For the case of New Airport Line in Beijing,the two proposed heuristic algorithms are capable of search for and producing better solutions,whose profits are 42,980 RMB and 43840 RMB,within 24 hours and 0.33 hours,respectively.These optimized plans are much better than the one produced by using the linear solver directly for 72 hours.