Home | DTI | 2007–08 funded proposals | Jon Weissman, Abhishek Chandra, Matthew Martimo, Mark Filipi
Jon Weissman, Abhishek Chandra, Matthew Martimo, Mark Filipi
Putting Next Generation Traffic Models on the Open Grid
Traffic modeling is a technique for exposing traffic and travel patterns in large urban areas. Understanding these patterns is crucial in urban development, transportation, city planning, and emergency response systems. However, currently deployed traffic models are limited in their complexity by the computational resources available to run them, resulting in their failure to uncover several desired patterns. In this project, we propose to host next-generation traffic models in Open Grid environments: computational platforms that are characterized by the availability of user-donated computing resources such as CPU cycles, storage, and network bandwidth for running collaborative applications. These environments are attractive because of their low cost and inherent scalability, and could be used to provide relatively cheap idle resources for compute-intensive applications. However, hosting a large-scale application such as traffic modeling in such environments is challenging because of the heterogeneity and unreliability of the underlying resources. In this project, we plan to develop techniques for achieving predictable performance from such environments, and demonstrate their efficacy by running a real-world large-scale traffic modeling application. We will conduct this research as a partnership between UMn/DTC, CitiLabs Inc., the developer of a widely-used traffic modeling application, and the Metropolitan Council/MnDOT, users of this software. The techniques developed in this project are expected to lead to future funding opportunities from federal agencies including NSF, DoE, and DHS, and state agencies such as MnDOT.
Digital Technology Center
499 Walter Library, 117 Pleasant Street SE, Minneapolis, MN 55455
P: (612) 624-9510