A methodology to characterise simulation models for discovery and composition: a system theory-based approach to model curation for integration and reuse

Abstract

We make a strong assumption that we can extract the necessary and sufficient information needed to decide whether a simulation model is suitable to an analyst's problem and that it contains information on how to be configured, integrated, and executed. This assumption allows us to review an approach that is founded on the theory of modelling and simulation and employs tools to develop, simulate, and apply models expressed in the discrete event system specification (DEVS) formalism, a sound systems theory-based computational methodology for system-of-systems model development and deployment. This foundation allows us to discuss how to apply the well-developed theory to curate simulation models so that they can more easily be discovered from a model repository given analytical objectives. After outlining the methodology, we place it in a broader context in which web and cloud-based technologies are employed for integration and to enable modelling and simulation as a service. Significant technical challenges are described that require continued research and development.