Data Farming from Simulations for Instantaneous Trade-off Analysis using Adaptive Batch-sequential Experiment Design

Our work will provide automated, adaptive, batch-sequential simulation methods for gathering data from a large, complex simulation model and developing metamodels that can be used for nearly instantaneous estimation of the simulation output at any set of inputs. Since most complex simulations take hours to run, these metamodels will allow for nearly instantaneous trade-off analysis. The results will greatly expand the ability of analysts and decision-makers to leverage the benefits of large-scale, complex simulation models in tactical situations.