Abstract

The reliability of transit modes is a crucial issue to guarantee their competitiveness and attractiveness. Experience shows that buses covering busy lines with short headways tend to arrive irregularly at their stops, often in bunches and therefore remarkably affecting the service. Hence, this work presents a stochastic model capable of representing the operational activity of a bus transit network and the service disruptions which eventually end up causing deviations from desired schedule. At the same time, the model is tested on an ideal route as a decision making tool.

Abstract

Service unreliability is widely recognized as one of the main deterrents for travelers to use buses as their primary mode of transportation. Bus systems are exposed to an adverse feedback loop that generates a tendency for buses to fall out of sync. This tendency can be counteracted by the application of control strategies that regulate the motion of the buses. Even though this is well known among transit operators and research has been devoted to solve this issue, in practice most bus systems rely on static control methods that ignore the real conditions of the system. Unfortunately, such an approach slows buses down excessively and it cannot compensate for severe disruptions.This research proposes a control method that applies dynamic holding based on real-time conditions, in order to synchronize the buses' motion and improve their service reliability without imposing an excessive burden on their speed. A formulation that improves and simplifies existing dynamic holding control strategies for isolated bus lines is developed. Stability conditions both for headway and schedule deviations are then derived, and the theoretical results are validated with a simulation case study. The improved formulation is then extended to corridors where multiple bus lines overlap, reproducing the stability results found in the single line scenario. A framework for a real application entailing a distributed system of on-board devices is also presented. The control performance is then validated, now using a real-world case study. Finally, the robustness of this control framework is assessed considering the effects of possible device failures. In particular, the effects of GPS malfunctions on the on-board devices are studied analytically, revealing the resilience of the proposed control strategy.