Reducing the validation bottleneck with a knowledge-based, distributed simulation environment

This paper reviews the concepts of a theory of modelling and simulation that relate to the validation enterprise. The theory provides a vocabulary, concepts, and mathematically rigorous tools with which to tackle problems in simulation model validation. We have implemented a hierarchical, distributed, object-oriented, and knowledge-based modelling and simulation environment in Ada. The environment, DEVS-Ada, provides portability, a standard model specification language, the means to manage a model repository, the ability to reuse models, and distributed simulation. We show how DEVS-Ada can exploit the parallelism intrinsic in the multiple execution of simultaneous experiments required for model validation. Faster execution of such a computationally intensive process reduces the bottleneck that validation imposes in the model development process and enables greater confidence to be achieved in the results. We also discuss the desirability of a more global view of validation which requires parallel symbolic analysis of a newly created model relative to existing models in a model base.