The Robustness And Control Strategies Of Bus Network Schedule Coordination With Uncertainty

With the gradual development of infrastructures and networks of public transport, the connectivity of lines is increasingly strengthened. Meanwhile, the travel needs are evolving and travelers expect seamless mobility solutions on the go, which requires more cooperation among different lines. Bus schedule coordination, which is an important step of timetable generation, aims at enhancing the systematic connectivity, reducing the travel time and improving the level of service(LOS). The bus network consists of transfer nodes and lines. The topologies of the networks, as well as each element in the networks, affect the efficiency of bus schedule coordination scheme(BSCS).Unlike railway transport, buses operate in the large open-ended environment, in which exists large amount of uncertainty that prevents vehicles from operating normally. It is difficult to maintain vehicles adherence to the basis plan in the uncertain environment, which hinders full realization of the efficiency of BSCS. In view of this, we propose the concept of BSCS robustness. From the view of ‘node, line, network’, we comprehensively explore how to improve the robustness of BSCS with uncertainty, and propose the corresponding control strategies with the idea of integrated disruption management. The main works and contributions are presented as follows:(a) By integration of slacks and holding, departure delayed control strategy is proposed. In this way, the uncertainty in travel time is addressed both in the planning and operation levels. Three scheduling modes(i.e., uncoordinated, departure punctual control and departure delayed control) are compared in two typical bus networks. It is found that there exists an interaction between the slack times and the safety control margin, which demonstrates that by including the real-time control strategy(i.e. the safety control margin setting) in the planning stage of setting the slack times, more cost-effective timetables with smaller slack times can be achieved.(b) Based on the empirical study, the concept of truncated delay distribution is introduced. The analytical expressions of systematic costs and transfer failure rates are derived. The impact of delay range on the optimal slack times and the impact of bus network structures on the BSCS robustness are discussed. It is confirmed that, through BSCS, the LOS of trunkand-feeder network can be improved while reducing the operating costs, and that the robustness is greater than that of loop network.(c) On the basis of Chapter 3, we explore how to further improve the benefit of BSCS in the trunk-and-feeder network by operation modes option. For simplification, the continuous approximation approach is employed. The critical condition of the best operation mode is analyzed. The co-optimization problem of zone and schedule coordination is then enhanced by inducing the travel time uncertainty. The results show that in low demand area, the demand-responsive transit(DRT) performs better than the fix-route transit(FRT), and that DRT allows lesser zones as such improve the efficiency, and that if well designed, DRT can improve LOS while reducing operating costs.(d) With consideration of incomplete schedule recovery and the resulting delay propagation effect, a model of real-time holding control at a transfer node is developed, which focuses on the uncertainty of forecasted delay time. The systematic analytical expressions are derived from the network-wide perspective, considering the effect of the holding control on the downstream of the line. And then the structural properties are demonstrated. The boundary conditions of both non-schedule recovery and incomplete schedule recovery are given, and the indicators of holding control robustness are proposed. It is found that the injection of schedule recovery benefits the holding control robustness, and that as the upstream transfer demand increases, the schedule recovery brings Pareto effect, meaning that both user and operator costs diminished.(e) From the planning level, by using discrete-event based simulation, we further investigate how to improve the stability of bus platoons in a bus corridor by optimal stop layout, as such reduce delay occurrence actively. The many-to-one(MTO) network is expanded to a more general many-to-many(MTM) network, in the sense that more factors can be incorporated into the bus bunching model, such as the alighting process, vehicle capacity and leftover passengers. We propose the system stability determinant method based on the time-spatial distribution figure of leftover passengers. We find that the spatially heterogeneous demand has a double-side effect on the bus system stability, which helps regularity recovery when the demand spatial distribution satisfies a certain condition.