Heuristic Estimation and Correlation of Statistically Significant Simulations

Physical Optics Corporation (POC) proposes to develop a new Heuristic Estimation and Correlation of Statistically Significant Simulations (HEUCOSS) system. This proposed software system is based on a new design that couples a range of high-performance modeling algorithms with artificial neural networks to assess the accuracy of the models. The innovation in heuristics, as well as accurate modeling and classification, will enable the system to create streamlined models of complex simulations and then to judge the quality of those models in terms of both accuracy and their performance. In Phase I, POC will demonstrate the feasibility of the HEUCOSS system by fully developing the prototype software, demonstrating its accuracy against known reference simulation data, and benchmarking its computational performance. In Phase II, POC plans to further develop the system by beginning small integrations with larger simulations in order to identify the aspects of the current models that will realize the greatest performance benefit from computational simplification with the least impact to accuracy. Approved for Public Release 14-MDA-8047 (14 Nov 14)