Transit network design for areas with low and/or heterogeneous demand

Low density and spatial heterogeneity in transit demand impose considerable challenges to both transit riders and service agencies. For example, lower demand forces transit agencies to provide sparser and less accessible service so as to stay economical, which however further deteriorates passenger experience and deters patronage. Spatially heterogeneous demand as well as city street characteristics (e.g., network layout) calls for variation in transit service, which often leads to higher system costs and undesirable passenger experience (e.g., transfers) as well. This dissertation proposes a series of transit network design methods that can be used to improve transit service under these circumstances. It first presents an alternative flexible-route transit system for low demand areas, in which each bus is allowed to travel across a predetermined area to serve passengers, while these bus service areas collectively form a hybrid “grand” structure that resembles hub-and-spoke and grid networks. We analyze the agency and user cost components of this proposed system in idealized square cities and seek the optimum network layout, service area of each bus, and bus headway to minimize the total system cost. We compare the performance of the proposed transit system with that of other conventional systems (e.g., fixed-route transit network and taxi service), and show which system is advantageous under different passenger demand levels. It is found that under low-to-moderate demand levels, the proposed flexible-route system has the lowest overall system cost. In the second part of this dissertation, a methodological framework is developed so that continuum approximation (CA) techniques can be used to design bus networks for cities where travel demand varies gradually over space. The bus-route configurations consist of (i) a main, city-wide grid with relatively large physical spacings between its parallel routes and the stops along those routes; together with (ii) one or more local grids with more closely-spaced routes and stops that serve neighborhoods of higher-demand densities. The so-called power-of-two concept is borrowed from the field of inventory control, and is enforced so that the local grids can be embedded seamlessly within the main one. Numerical experiments illustrate the value of the resulting heterogeneous route configurations, which have the potential to reduce the costs to both the bus users and the operating agency, as compared against the costs of the optimal homogeneous bus-route grids. Differences of about 5-10% are observed for a set of numerical examples that cover a wide range of demand distribution patterns. Most of the savings are due to the diminished access costs that users incur when high-demand neighborhoods are served by local grids with closely-spaced routes and stops. The same CA methodology is used to design a simplified grid network system with variable bus spacings that can address spatially heterogeneous demand. The continuum approximation enables us to estimate the cost components of the system locally and design the system layout accordingly. The simplified grid system varies the bus line spacings, which makes the network more responsive to the varying demand. It saves user costs in relatively high demand areas and saves agency costs in lower demand areas. Numerical results show that the design can improve the total cost of the system by between 4% and 6% as compared with traditional grid designs under the chosen demand distribution patterns. We further extend the network design framework from grid city street networks to a radial one, where buses can travel either radially (from center to outer part and vice versa) or circularly (clockwise and counterclockwise along rings). Using the CA method, we analyze user and agency costs and propose a design framework that minimizes the total system cost. The optimum design provides the bus spacings of the radial and circular lines and the bus headway. Numerical results show that the proposed method can be useful to design an efficient system when the city streets form a ring-radial network. The last part of this dissertation is devoted to real-world applications of the proposed design methods. We apply the flexible-route transit network to satisfying evening/night demand in the campus area of Urbana-Champaign, which is currently covered by a manually dispatched “Safe-Ride” system. The new transit system is shown to outperform the current system and achieve several improvements. We also design a new transit network for the City of Weihai in China, where the network includes both grid and polar network components to best serve the city’s geographical layout. Numerical experiments also give insights on the influence of key decision variables such as bus line spacing, stop spacing, and headway under different design schemes.