Transportation networks are inherently uncertain due to random disruptions; meanwhile, real-time information potentially helps travelers adapt to realized traffic conditions and make better route choices under such disruptions. Modeling adaptive route choice behavior is essential in evaluating Advanced Traveler Information Systems (ATIS) and related policies to better provide travelers with real-time information. This dissertation contributes to the state of the art by estimating the first latent-class routing policy choice model using revealed preference (RP) data and providing efficient computer algorithms for routing policy choice set generation. A routing policy is defined as a decision rule applied at each link that maps possible realized traffic conditions to decisions on the link to take next. It represents a traveler's ability to look ahead in order to incorporate real-time information not yet available at the time of decision. A case study is conducted in Stockholm, Sweden and data for the stochastic time-dependent network are generated from hired taxi Global Positioning System (GPS) traces through the methods of map-matching and non-parametric link travel time estimation. A latent-class Policy Size Logit model is specified with two additional layers of latency in the measurement equation. The two latent classes of travelers are policy users who follow routing policies and path users who follow fixed paths. For the measurement equation of the policy user class, the choice of a routing policy is latent and only its realized path on a given day can be observed. Furthermore, when GPS traces have relatively long gaps between consecutive readings, the realized path cannot be uniquely identified. Routing policy choice set generation is based on the generalization of path choice set generation methods, and utilizes efficient implementation of an optimal routing policy (ORP) algorithm based on the two-queue data structure for label correcting. Systematic evaluation of the algorithm in random networks as well as in two large scale real-life networks is conducted. The generated choice sets are evaluated based on coverage and adaptiveness. Coverage is the percentage of observed trips included in the generated choice sets based on a certain threshold of overlapping between observed and generated routes, and adaptiveness represents the capability of a routing policy to be realized as different paths over different days. It is shown that using a combination of methods yields satisfactory coverage of 91.2%. Outlier analyses are then carried out for unmatching trips in choice set generation. The coverage achieves 95% for 100% threshold after correcting GPS errors and breaking up trips with intermediate stops, and further achieves 100% for 90% threshold. The latent-class routing policy choice model is estimated against observed GPS traces based on the three different sample sizes resulting from coverage improvement, and the estimates appear consistent across different sample sizes. Estimation results show the policy user class probability increases with trip length, and the latent-class routing policy choice model fits the data better than a single-class path choice model or routing policy choice model. This suggests that travelers are heterogeneous in terms of their ability and willingness to plan ahead and utilize real-time information. Therefore, a fixed path model as commonly used in the literature may lose explanatory power due to its simplified assumptions on network stochasticity and travelers' utilization of real-time information.