Numerous transportation analysis and modeling tools have been developed over the last 30 years, and the data collection for these models was a personnel-intensive process requiring significant resources. As a result, there typically were insufficient amounts of data to verify transportation models under diverse operating conditions. However, with the proliferation of sensors to support intelligent transportation systems, researchers now find themselves in a data glut and an information shortage. With this data overload, there is a need to shift the modeling focus toward a state-based modeling approach. Fundamental to this is the need to explore the question of what constitutes a parsimonious state definition that effectively captures the dynamics of a large-scale, regional transportation system. Discussed are a proposed framework for a state-based modeling paradigm and, specifically, two key concepts to this new approach: case-based reasoning and simulation for model generation.