Understanding and modeling route choice decisions

点击翻译按钮获取中文摘要

Although travel behavior research in general has advanced considerably over the past decade, spearheaded by new data collection technologies and new modeling methods, one area that has seen less work is that pertaining to route choice. In part, this stems from a lack of data pertaining to observed routes - data that can now be collected via GPS technology. At the same time, in an era characterized by rising oil prices, energy conservation, traffic congestion, aging populations and climate change, there is a growing need to improve our understanding of the spatial manifestation of activity/travel behavior - that is, navigating urban transportation networks. Since travel demand forecasting is done using some form of traffic assignment algorithm, the research proposed can reveal whether the assumptions underlying such algorithms are valid. If not, then one must question the validity of predictions from such models - predictions that are used to guide development in urban areas. In other words, how can a transportation planner enhance quality of life when the model outputs are simply inaccurate? The goal of the proposed research is to understand and model route choice decisions. To this end, a comprehensive GPS-enhanced data set collected in 2007-08 for a sample of households in the Halifax Regional Municipality, Nova Scotia, Canada is used to address the following objectives: (1) to compare alternative methods for generating route choice sets; (2) to compare route choice models using alternative modeling structures that have been proposed to capture similarities between routes; (3) to examine the effect of choice set reduction on model estimates; (4) to determine if route choice behavior varies between groups of individuals based on their socio-demographic characteristics; (5) to determine if route choice behavior varies by trip purpose; and (6) to determine if route choice behavior varies by time of day.