Transcript Slide 1
Basics requirements Spatial considerations (resolution) Input Maps -- (raster images) Veg/Elevation/TSI/WMA/RD) Initialization (on computational domain) 1 vegetation classification 2 GWWA HSI 3 CERW HSI Output Maps map layers (raster images) statistical analysis on raster images Computational Science for NRM D. Wang Basics requirements Temporary considerations Simulation period (30 years?) Timestep (1 year) Computational Science for NRM D. Wang Basics requirements Control Variables/Scenarios 1) total_area 2) seed_location 3) site_extension 4) min_max_size Optimization criteria 1) projected_mine_area 2) max_HSI_area Computational Science for NRM D. Wang Implementations How to represent maps? - array - customized template (TIEM’s landscape library) How to represent time? - integer - customized time library (TIEM’s Date library) Computational Science for NRM D. Wang Functionalities Aging (very clear) Mining (complicated) input (TSI, Total_area, max_min_size) site_growth *** (core function) output (a series of maps/tables) Computational Science for NRM D. Wang Model Structure Most Simple cases (fixed scenarios) 1) we know when to start and stop 2) we have no need to learn during the simulation 3) present the results at end of simulation Mining options are limited Growth function is given (we already know the output of mining function Computational Science for NRM D. Wang Basic Data Structure Arrays: Veg/Elevation/TSI/WMA/RD (read-in) vegetation classification time_since_mining GWWA_HSI CERW_HSI Computational Science for NRM D. Wang Data Flow Model Initialization Loop 1: From first option to the last option Loop 2: From first year to the last year 1) Get the scenario map (sm) (read-in) 2) Compute and save the HSI maps End of Loop 2 Analyze those HSI map, save results End of Loop 1 Model Finalization Computational Science for NRM D. Wang Optimal Control for Given Scenarios Which scenarios is better? 5% 10% 20% Possible procedure Loop 1: From scenarios 1 to 3 Loop 2: From first year to the last year 1) Compute the scenario map for given scenario (how many? Complete enumeration?) 2) Compute and save the HSI maps End of Loop 2 Analyze those HSI map, save the best/better results End of Loop 1 Computational Science for NRM D. Wang Difficulties Associated with Optimal Control What is the best scenarios? definition Is this age-based HSI model reliable? Which control variables, criteria are good candidates for this kind of control? … IF All of those mentioned above are clear, is this approach computational feasible? More important, we need feedbacks from all partners (computer scientists, and ecologists, and natural resource managers …) (interactive GIS-based modeling?) Computational Science for NRM D. Wang Questions on implementation details… Language (you chose)– native language please Computational platform Data structures for easy storage information handling One functionality (aging) Map formats (binary, ASCII) Integration with GIS for exporting and importing maps (non-interactive) Performance issues … Computational Science for NRM D. Wang