Accessibility measures are necessary for evaluating the benefits of proposed transportation improvements. However, they often do not account for travel time unreliability, but instead incorporate deterministic and time-invariant travel times. This approach risks mischaracterizing the accessibility experienced by travelers. In this paper, we review recent literature on accessibility and travel time reliability with a focus on transit and introduce an approach to joint accessibility-reliability measurement that relies on a behavioral perspective. Using this behavioral perspective, we propose that existing accessibility measures be implemented using travelers’ total travel time budget as a measure of travel time, and that varying departure time strategies depending on service characteristics be considered. The total travel time budget can reasonably be quantified with a high percentile of the total travel time distribution. However, we note that different percentiles may be more appropriate for different traveler types, as these percentiles correspond to varying tolerances for late arrivals. This behavioral perspective can be operationalized with commonly used accessibility measures, such as the cumulative opportunity measure, and with real-time vehicle location data. We include a demonstration of the potential changes in accessibility estimates when accounting for travel time unreliability, with a simplified case study of a transit route in San Francisco. The results show a considerable reduction of the number of opportunities available to travelers when the calculation is based on the latter—between 5.9% and 37.9% less, depending on various factors. Such differences have the potential to significantly affect the accessibility benefits of transit capital investments.

无障碍措施对于评估提议的交通改善措施的好处是必要的。但是，它们通常不考虑行程时间的不可靠性，而是合并确定性和时不变的行程时间。这种方法冒着误以为旅行者所经历的可访问性存在风险。在本文中，我们以交通为重点，回顾了有关可达性和旅行时间可靠性的最新文献，并介绍了一种基于行为观点的联合可达性-可靠性测量方法。使用这种行为观点，我们建议使用旅行者的总旅行时间预算作为旅行时间的措施，以实施现有的无障碍措施，并考虑根据服务特点而变化的出发时间策略。总旅行时间预算可以合理地量化，且总旅行时间分布的百分比很高。但是，我们注意到，不同的百分位数可能更适合于不同类型的旅行者，因为这些百分位数对应于不同的迟到公差。可以使用常用的可访问性度量（例如累积机会度量）和实时车辆位置数据来实现这种行为观点。我们以旧金山过境路线的简化案例研究为例，说明了考虑到旅行时间的不可靠性时可访问性估计值的潜在变化。结果表明，根据旅行者的计算，旅行者可以利用的机会数量将大大减少-减少5.9％至37.9％，取决于各种因素。这种差异可能会严重影响过境资本投资的可及性收益。