Optimizations Of Route-level And Network-level Stop Spacings And Facility Configurations Of Urban Public Transport

With the rapid development of urbanization process,the existing road resources and transportation facilities become difficult to adapt to the high-speed increase of population and accompanying motorized travel demand. Public transport has pronounced advantages in the per capita road traffic resources utilization efficiency, energy conservation, exhaust emissions reduction and many other aspects as compared with other motorized travel modes. To give priority to the development of public transport is a sustainable measure aiming at alleviating traffic congestion. Configurations of route-level and network-level spacings of public transport are pivotal factors that determine service capability, operational efficiency and level of service. Thus, it is of great significance to study the optimization of stop spacings at both route and nework levels and the coupling between transit supply and travel demand distribution in order to ahieve the enhancement of operational efficiency of urban public transport system.The study tries to explore and analyze the optimization of stop spacings of urban public transportation system at both route and nework levels. The work was sponsored by the Key Project of National Natural Science Foundation of China "Basic Theory and Key Technology of Efficiency Improvement for Multi-modal Public Transport System in Modern City"(No.51338003). Considering land usage, the study first propsed a bi-level optimimation model to analyze the setting of physical stop spacing at the route level. Then the study proposed two mathematical models to optimize the design of limited-service services and corresponding operational stop spacing at the route level based on temporal and spatial variations in travel demand. From the perspective of transit nework, the study proposed the continuum approximation model to investiage the coupling between transit network configuration and demand distribution. Later, the study proposed discrete network-based optimization model to obtain the optimal stop spacing and associated stop locations. At last, the study discussed how to improve the design and layout of feeder facilities as well as transfer services exclusive for the weak service areas. Results provide theoretical basis for the planning and operation of urban public transport system. More specifically, this study is separated into the following contents:First, transit demand distribution is closely related with the land use conditions, and the setting of physical stop spacing should cater for the basic travel demand along a transit route.The study porposes an approach to optimize physical stop spacing at the route level based on land use. Hierarchical clustering is used to analyze the function of various land use types for existing stops and then separate the service area of a transit route into several service catchments. Later, a bi-level optimization model is proposed to analyze the optimal stop spacing for these catchments with different land use types. The upper-level model minimizes a cost function which includes the benefits of the local authority, passengers, and operators,and the lower-level model is associated with transit assignment. Opitmal solutions determine the number of stops in each catchment and near-optimal solutions guide for the optimization and adjustment of transit stops.Second, once physical stop spacing is optimized, conducting limited-stop service is a good option to better adapt to temporal and spatial variations in travel demand along the transit route. The study analyzes two typical approaches to optimize limited-stop service and corresponding operational stop spacing along a transit route. The first one proposes an optimization model to optimize limited-stop service and corresponding operational stop spacing along a single transit route. To better reflect the reality, the model considers stochastic travel time, vehicle capacity and in-vehicle congestion. The study proposes a hybrid artificial bee colony and Monte Carlo method to solve the model. The second one proposes an optimization model to optimize limited-stop service and corresponding operational stop spacing along two parallel transit routes. Transit vehicles in one route dwell at only a few stops along the route while vehicles in the other parallel route serve all the intermediate stops.The model considers vehicle capacity constraint, service frequency constraint and limited-stop service constraint. The study proposes a multi-start iterated local search method to solve the model.Third,based on the city's spatial and temporal demand distribution, the study seeks the relationship between transit supply and travel demand, and analyzes the systemic coulpling characteristics between transit network-level stop spacing and travel demand distribution.From the perspective of transit network, the study proposes two coupling models. The first one is two continuum approximation models: the former one is a static continuum approximation model based on physical stop spacing, and the latter one is a dynamic continuum approximation model based on operational stop spacing. The study uses analytic methods to solve and compare the results of two continuum approximation models. Then, the study perfoms a sensitivity analysis on vehicle capacity, fleet size, headway and other parameters, and analyzes their effects on the transit network configurations. The second one proposes a discrete optimization model to select appropriate stops from a set of candidate pickup locations in urban area. The constraints in the model includes access coverage which varies with land use and the effect of transit stop congestion on road traffic flow. The study uses CPLEX to sovel the model.Finally,the service area that both route-level and network level stop spacings are optimized is the city's central district, whereas the suburban area becomes a weak link for the urban public transport system. Based on the optimized route-level and network level stop spacings, the study analyzes two important feedering facilities to connect the suburban area and transit hubs/stations within the central district. The first one is the design of suburban feeder transit route. The study proposes a special model and a generic model. Two models are solved by a dynamic programming approach and an artificial bee colony approach respectively. The second one is the dermination of the optimal layout design for public bicycle system within the attractive scope of a transit hub/station. The study proposes a hierarchical layout approach of public bicycles and associated rental stations. Then,the study proposes an improved immune algorithm to determine the optimal number and locations of service stations.The study provides beneficial theoretical support and reference for the planning and design of transit route-level and network-level stop spacing and facility configurations in urban areas.