Model Level Integrated Simulation Architecture for Collaborative Development

We propose to develop a simulation architecture framework that enables development teams to integrate algorithmic models into a simulation solely through a collaboration layer/interface and software controls. The proposed architecture will address: ease of model integration, model composability, execution speed, and ease of data collection and retrieval. To achieve these objectives, we propose that the Discrete Event System Specification (DEVS) modeling and simulation formalism provides the theoretical and practical basis to manage whole ecosystems of related models through a simulation architecture framework. DEVS environments support many of the sub-objectives; however, they do not explicitly support the integration of algorithmic models through interfaces. Algorithmic models can be incorporated into DEVS atomic models providing access to simulation constructs defined by DEVS simulators. Crucially, we distinguish three distinct types of users to be supported: model developers (may create atomic models, modify existing models to create model variants, or compose coupled models), subject matter experts (may compose coupled models selected from a model library through a convenient interface not requiring programming experience), and production users (may simulate models on a variety of execution platforms). Approved for Public Release | 18-MDA-9817 (23 Oct 18)