P10: A Grid-enabled Simulation-Optimization Framework for Environmental Characterization Problems Session: Posters: Grid Event Type: Poster Time: 5:00pm - 7:00pm Speaker(s): Baha Y. Mirghani, Michael Tryby, Ranji Ranjithan, Kumar Mahinthakumar, Derek Baessler, Nicholas Karonis Location: 4th Floor South Lobby Abstract: Environmental characterization involves the resolution of unknown system characteristics from observation data, and thus can be categorized as an inverse problem. One solution approach for inverse problems is through coupling of environmental simulation models with global search methods. Computational tractability of such a simulation-optimization approach could be enhanced through a grid-computing framework. This poster describes the development of the LArge Scale Simulation Optimization (LASSO) framework, which utilizes standard communications protocols and the MPI2 API to establish a connection between a centralized search application with simulation models running on TeraGrid resources. We report our experiences in developing this grid-enabled framework, and demonstrate preliminary and enhanced set of results for a groundwater release history reconstruction problem where significant performance improvements are observed. The poster will show details of the application problem, solution approach, the LASSO architecture diagrams on some TeraGrid sites, and performance results for single-site and cross-site runs on the TeraGrid. Chair/Speaker Details: Baha Y. Mirghani North Carolina State University Michael Tryby North Carolina State University Ranji Ranjithan North Carolina State University Kumar Mahinthakumar North Carolina State University Derek Baessler Northern Illinois University Nicholas Karonis Northern Illinois University