Data organization and data comparison for model validation in faster-than-real-time simulation

Validation is an essential feature of faster-than-real-simulation (FRTS), since models must be validated prior used for near-future predictions. As no relevant approaches exist in the literature, the paper contribution is to propose an approach a) accomplishing validation as a real-time, automated process, with low time overhead, b) dealing with the complexity of validation data, as the number and type of the data compared may be changing and c) realizing the transition from validation process conceptual design to an efficient real-time execution. A formal data organization scheme (data model) and algorithms for constructing and accessing it are implemented for this purpose. Realization of model validation in a FRTS experiment on a single-queue/multi-server processing system is presented to exhibit the applicability of the proposed approach. The proposed organization scheme may also be applied in simulation studies where timing and automation requirements are not critical.