Layout Optimization Of The Feeder Bus Routes Connecting Subways At Bus Stops

Based the features of large volume, running time reliability, high delivery efficiency, energy saving and environmental protection, urban rail transit is becoming the important solution to alleviating the traffic pressure. Efficient of the urban rail transit system depends on the cooperation with others modes of transportation, especially the feeder bus.Aiming at the unbalanced utilization rate of bus stops, the low level of passenger service and the long waiting time due to unreasonable distribution of feeder bus lines, an optimization model is established to obtain the shortest total passenger’s transfer distance and the minimum the total time of all buses queuing to dwell at bus stops under the limit of the throughput capacity of platform. This paper gives Pareto frontiers by ε constraint approach and employs weight coefficients to indicate the preference of decision-makers. Case studies of the actual stop are implemented to verify the validity and sensitivity analyses are conducted. The paper includes the following main parts:(1) Analysis on factors of berth of bus vehicles and the arrangement of bus lines is the most fundamental part in this paper. Different locations and types of bus stop have a different influence on road traffic flow. Considering characteristics of feeder bus stops, which comprise more line buses, more than one platform, longer bus stopping time for passengers alight or egress and larger ridership, bus stops located around intersection and harbor-style bus stops are more appropriate for layouts of subway station and bus stop.(2) The shortest transfer distance of total passenger is the one of objects. The distances from every bus berth to the subway entrance are different at the same platform, which are smaller by comparing with the distance from the bus stop to subway entrance. Furthermore, passenger transfer between bus and bus are more complicate than transfer between subway and bus while having no significant impact on transfer between subway and bus. This paper calculates the transfer distance by mainly considering transfer passenger between subway and bus.(3) The minimum bus vehicles waiting to queue to draw up at berths for all bus platforms is another object. The vehicle arrival distribution for multiple bus routes complies with Poisson distribution and service time of bus, which is dwelling time spent in bus stops, follows Erlang distribution. Considering the bus cross among different bus platforms non-significant, the model achieves the total time of all buses queuing to dwell at bus stops by accumulation bus waiting time in each bus platform using M/M/C model in queuing theory.(4)This paper gives Pareto frontiers by ε constraint approach and employs weight coefficients to indicate the preferences of decision-makers. This model using ε constraint approach to convert multi-objective model into single objective is a nonlinear model, which can’t be solved by simplex method. However,0-1integer programming can only be applied to integer programming. This paper improves0-1programming to solve nonlinear integer models and uses Lingo software to obtain Pareto frontier. Finally this paper uses distance based approach to achieve the optimization alternative.(5)Case studies of Xiyuan station of line4of Bejing subway are implemented to verify the validity and practicability of the proposed methods and a sensitivity analysis on all alternatives are conducted, whose reaction to the weight coefficient are different. Alternatives are sensitive when two object standard values exit great difference, which results in wider gap on total object values under different weight coefficients.