Real-time optimization of passenger collection for commuter rail systems.

PDFTech ReportSWUTC/14/600451-00082-1DTRT12-G-UTC06Railroad commuter serviceMode choiceFeeder busesReal time informationBus stopsBus routesOptimizationTabu searchMixed integer programmingPublic transitTexasTexas A&M Transportation InstituteYu, YaoMachemehl, RandyHakimi, ShadiUniversity of Texas at Austin. Center for Transportation ResearchSouthwest Region University Transportation Center (U.S.)United States. Dept. of Transportation. Research and Innovative Technology AdministrationUS Transportation CollectionCommuter rail systems are being introduced into many urban areas as an alternative mode to automobilesfor commuting trips. The shift from the auto mode to rail mode is anticipated to greatly help alleviatetraffic congestion in urban road networks. However, the right-of-way of many existing commuter railsystems is usually not ideally located. Since the locations of rail systems were typically chosen long agoto serve the needs of freight customers, the majority of current commuter rail passengers have to take anon-walkable connecting trip to reach their final destinations after departing even the most convenientlylocated rail stations. To make rail a more viable, competitive commuting option, a bus feeder or circulatorsystem is proposed for seamlessly transporting passengers from their departing rail stations to final workdestinations. The primary research challenge in modeling such a bus circulator system is to optimallydetermine a bus route and stop sequence for each circulating tour using the real-time demand information.In this paper, we termed this joint routing and stop optimization problem the circulator service networkdesign problem, the objective of which is to minimize the total tour cost incurred by bus passengers andoperators while minimizing the walk time of each individual bus passenger. A bi-level nonlinear mixedinteger programming model was constructed and a tabu search method with different local searchstrategies and neighborhood evaluation methods was then developed to tackle the circulator servicenetwork design problem