The objective of this Faculty Early Career Development (CAREER) program award is to investigate the relationships between transportation networks of different scales/sizes, learn how to quantify these relationships, and discover the implications for transportation planning. The research focuses on three examples where scale and size are important issues: freight models where a multistate or national scale is considered in addition to a local scale, allocation of transportation funds within a region, and transportation networks that change rapidly with time (e.g., due to accidents or other disruptions). Current multi-scale models are often constructed in an ad hoc manner, without understanding the fundamental connections between the models. The research investigation in this grant will be guided by the conjecture that "mixing" models incorporating simplified versions of one model into another, rather than separating them completely is the key to efficiently finding consistent solutions.
If successful, this work will allow network models to represent much larger regions, specifically supporting improved transportation planning, but also contributing to multi-scale network models in other domains. A complementary education and outreach plan to the research plan has four major components. First, active-learning curricula will be developed for transportation engineering and network analysis courses, building on recent pedagogical research. Second, a library of state-of-the-art transportation algorithm codes will be made public, allowing students to work with large-scale practical data in course projects and allowing practitioners to experiment with state-of-the-art models, facilitating technology transfer. Both the code and curriculum materials will be released in open-access form, free of charge. Third, the PI will actively recruit and mentor undergraduate and graduate researchers, interfacing with university programs to identify talented students from underrepresented groups. Finally, a new course in multi-scale modeling will be developed, based on the research results.
