A dynamic scheduling process and methodology using route deviations and synchronized passenger transfers for flexible feeder transit services. (October 2022)

• Record Type: Journal Article
• Authors: Yang, Liya; Shang, Pan; Yao, Yu; Zeng, Ziling
• Is Part Of: Computers & operations research. Volume 146(2022)
• Publication Date: 2022-10
• Subjects: Transportation -- Feeder transit services -- Synchronized passenger transfer -- Space–time–state network -- Coordinate descent-based solution framework; Operations research -- Periodicals; Electronic digital computers -- Periodicals; 004.05
• DOI: 10.1016/j.cor.2022.105917
• Languages: English
• ISSNs: 0305-0548

Abstract

Highlights: Proposed a dynamic scheduling methodology for flexible feeder transit services. Integrate route deviation and synchronized passenger transfer into scheduling process. Formulate a time-discretized multi-commodity network flow model. Construct a multi-dimension space–time–state network as solution space. Propose a dynamic programming-based coordinate descent framework to solve the model. Abstract: The general lack of first/last mile connectivity is one of the main challenges encountered by intercity transport systems. New transportation alternatives called flexible transit systems, which combine on-demand service adjustment capabilities to the regular route and master schedule characteristics of the conventional transit, are therefore increasingly proposed. Feeder transit services can run as one type of flexible transit services, connecting a service area to a major intercity transport network through a transfer hub. Considering the unstable and constantly changing characteristics of daily passenger demand, a dynamic scheduling methodology for feeder transit services is proposed to minimize passenger crowding in vehicles. Deviations in the base route are allowed for accommodating optional requests or visiting potential transfer locations, which further offer a new type of flexible scheduling operation called synchronized passenger transfer. Operators can dynamically expand the spatial coverage of vehicles out of their regular routes to meet each other at appointed locations and times. Then, passengers can transfer between vehicles to minimize or even avoid passenger crowding. In this study, a multi-commodity network flow optimization model with side constraints in a three-dimensional space–time–state network is proposed for this type of dynamic scheduling problem. Additionally, a coordinate descent-based solution framework consisting of three steps (augmentation, decomposition, and linearization) is developed. The proposed model and solution approach are investigated using an eight-node simple network and the Sioux-Falls network, and the Chicago sketch network was adopted to demonstrate the practicability of the proposed model.