Simulation Model Validation

The proposed program will implement a consistent, user-friendly, Microsoft Windows (C++) based framework for the "Simulation Model Validation". Software will be written in such a way as to be easily portable to UNIX class workstations. The Phase I objective will be to demonstrate the feasibility of the Ontar multivariate methods of model validation. Over the last 15 years Ontar has been working in the field of atmospherics, sensor and scene simulation. After many years of experimenting with validating models, we stand by the simple conclusion that a model is validated when the system under testing cannot tell the difference, in statistical sense, between the real and the simulated results. The multivariate technique of discriminate function analysis (DFA), is a formalized way to measure and quantify the difference between samples. for validation the sample types are test samples and real samples. Because of Ontar's background and experience, we will provide considerably more than what is called for in the Phase I solicitation. We regard this program as an opportunity to implement a complete "soup to nuts" Simulation Model validation" customer support system. We will provide additional services, e.g. on-line documentation, training, newsletter, and support. Ontar anticipates that the successful completion of this program will lead to a CRDA (Cooperative Research and Development Agreement) for the long-term support and development of "Simulation Model Validation". Ontar has successfully entered into agreements with several government agencies and other organizations for technology transfer. In an era of severe budget cuts this may be the best way to guarantee the life of the product.