Analysis And Optimal Design Of Urban Rail Transit Station Service Facilities

The urban rail transits are playing a significant role in the urban transport with the rapid progress in urbanization,especially in the metropolises.The urban rail transit stations are the operational systems consisting of a framework of infrastructures,service facilities,and personnel;they are the points of connection between arrivals and departures of passengers.In recent years,investment and improvement in the urban rail transits encouraged the people to switch from driving to transits.The performance of the urban rail transit station service facilities naturally became a great concern to both passengers and operators.The better performance of these service facilities is the reflection of enhanced design while the inadequate design often leads to high-level congestion,the longer travel time of passengers between the service facilities,inefficient space utilization,resource wastage and increase in the waiting time of passengers which in turn implies that there is a direct correlation between design and performance.However,at the same time,the urban rail transit station service facilities are obliged to hold the Level-of-Service(LOS)specified by the codes and design manuals,but they have several shortcomings,such as 1)They do not take into account the randomness in passengers’ arrival process;2)They do not consider the fluctuation and state-dependence of the passengers’ walking speed in the facilities;3)The service facilities are configured and designed separately and the correlation between them is ignored;4)They do not consider the blocking phenomenon or congestion that happens when the passengers’ traffic demand approximates to or exceeds the serviceability of the service facilities.Moreover,the existing analytical approaches are highly complicated and lead to challenging complications.The service facilities of the urban rail transit station and passengers’ flow comprise a kind of unique queuing system that may be modeled and analyzed by using analytical queuing and simulation models.Therefore,the study reported in this dissertation details the Discrete-Event Simulation(DES)as well as the simulation-based optimization approach for the analysis and design of urban rail transit station service facilities,considering both the randomness in passengers’arrival process and the service times of the facilities.The Phase-Type(PH)distribution considers the randomness factor and therefore it is used to fit the passengers’ arrival and service processes.Moreover,the simulation-based approach,integrating the PH-based DES models of the service facilities and the optimization algorithm based on Genetic Algorithm(GA)is used to design the facilities and eliminates the need to solve explicit analytical expressions over a large time span,as in the case of mathematical optimization.Three different service facilities of urban rail transit stations,including circulation facilities(corridors and stairs),Ticket Service Facilities(Ticket Vending Machines(TVM),Ticket Offices(TO)and Automatic Fare Gates(AFG))and the Elevator facility are studied in this research.First,these service facilities are described as queuing system and the analytical queuing models based on the PH distribution are developed to evaluate the performance of these service facilities.Then,the urban rail transit station service facilities are translated to the DES models with the passengers’arrival and service processes based on the Phase-type(PH)random variates.The elevator facility,ticket facilities and circulation facilities are modeled as PH/PH[K]/1/inf DES model,PH/PH/c/∞DES model and PH/PH(n)/C/C DES model,respectively.The PH-based analytical models serve as the benchmark to verify the accuracy of the PH-based DES models and therefore the DES models are first verified by analytical models and then the performance analysis are carried out.The results revealed that the models based on the exponential distribution(M)and deterministic distribution(D)always underestimate the performance measure values for the different passengers’arrival rates,service rate of service facilities and the Squared Coefficient of Variation(SCV).The existing models of these service facilities fail to consider the randomness in passengers’ arrival as well as randomness and state-dependence in service processes and therefore do not depict the actual conditions of the urban rail transit stations.Furthermore,the width of the circulation facilities(corridors and stairs)and configuration(number)of ticket facilities in the urban rail transit stations are the vital design parameters.To support width design for the circulation facilities network and optimal configuration of ticket facilities,the simulation-based optimization approach which fully considers the randomness and state-dependence in passenger flow as well as the Level of Service(LOS)and blocking probability control is proposed.The new approach provides automatic reconfiguration of the widths of circulation facilities and the ticket facilities by a concurrent implementation of a Discrete-Event Simulation(DES)models and the Genetic Algorithm(GA).The proposed simulation-based optimization approach and the existing design approaches based on the exponential distribution(M)and deterministic distribution(D)are applied to design the widths for the circulation facilities network as well as for the optimal configuration of ticket facilities.The results reveal some interesting findings:(1)The circulation facilities designed by the proposed approach has larger widths;(2)When the SCV of arrival interval grows for the circulation facilities,the widths of the proposed design approach increase while the widths of the other two approaches stay the same;(3)The widths obtained by the proposed approach increase faster than that of the other two approaches when the arrival rate rises;(4).The required number of ticket facilities increase with increase in passengers’ arrival rate as well as with the increase in SCV of arrival interval;(5)The optimal configuration of ticket facilities based on the existing models results in less number of ticket facilities because of neglecting the passengers’ randomness effect.The proposed PH-based DES models and PH-based simulation-based optimization approaches overcome the shortcomings in the existing analysis and design approaches by fully considering the randomness and state-dependence and therefore the squared coefficient of variation of arrival process as well as service processes cannot be ignored during the analysis and design of the service facilities.In addition to that,the PH/PH(n)/C/C DES model for circulation facilities network is reduced to a PH/PH(n)/C/C DES model for single corridor facility and the simulation-based optimization approach is used to design the width of the single corridor facility.This model can also serve as a tool to verify the accuracy of analytical model for corridor facility based on the PH distribution.The numerical experiments are conducted for the single corridor facility under the LOS ’B’,’C’ and ’D’ to compare the proposed model with the existing design approaches.It reveals that:(1)The width obtained by the proposed model for the single corridor is higher than the existing width design models;(2)When the SCV of passengers’ arrival interval increases,the single corridor width increases more for the proposed model than the existing design models;(3)The increase in the.length of the corridor has no significant effect on the corridor width.