Design Of Competitive And Hierarchical Transit Network With Grid And Hub-And-Spoke Structure

Transit systems are a key strategy in urban areas to relieve traffic congestion and promote the sustainable development. Bus transit plays an important role in urban transport because of the flexibility of route configuration, the easy construction, the low investment and operating costs. The effectiveness of the bus system is determined by the structure of the network essentially, thus an improvement in network topology are the main elements to redirect the bus transit situation and to enable a more efficient and competitive system.A great deal of research on bus network design problem has been carried out and the discrete models are widely used. The resulting networks are complex and consist of a set of independent routes. In contrast, analytical models take continuous and compact formulations based on simple network schemes, which makes studying the behavior and design of bus systems easy. Therefore, the analytical model is developed in this thesis to design bus networks for cities whose road structure presents a grid layout. The network consists of three levels of routes and shows a hybrid structure with a grid in the central square and hub-and-spoke in the periphery. The sum of the agency costs and the generalized user costs are included in the objective function to be minimized. The problem is solved by Genetic Algorithm (GA) developed with MATLAB. Then, the competitiveness of the optimal network is compared with that of the passenger car. Furthermore, a sensitivity analysis is carried out to foresee the model’s behavior.Firstly, the fundamental theory research is conducted. In order to construct a multi-level bus system, the basic network structure, the idea and design method of hierarchical bus network, and the rule of routes connection is analyzed. The three levels of bus system are determined, which consists of express route, routine route, and branch route. Besides, common model for bus network design and the component of total costs in the objective function are formulated. The solution algorithms for multivariate nonlinear model are compared. The GA toolbox based on MATLAB is chosen as the solution method.Secondly, the analytical model is developed and solved. The objective function is the minimization of the sum of agency costs and the user costs. Important agency costs include the construction and maintenance costs of infrastructure and operation costs. User costs are proportional to the travel time in the user transit chain, including access, waiting, riding in-vehicle, and transferring. The constraints of the decision variables and the vehicle capacity are considered. Moreover, the solving procedure of GA is coded in MATLAB. The decision variables, components of costs, and the level of service parameters in the optimal network are calculated.Finally, the analytical model is applied for the bus network design of the urban district of Xi’an. Based on the characteristics parameters of the urban, the users’ trip, and the bus operation, the optimal network structure is determined. With the optimal network, the competitiveness is compared with that of the passenger car. The components of costs and the level of service are analyzed. Different scenarios are designed to identify the effects of travel demand and urban scale on the network structure. Furthermore, sensitivity analysis with regard to the decision variables, demand distribution, and the transport modes are carried out to evaluate the stability of the developed structure.