In this paper, we review the literature related to the reuse of computer-based simulation models in the context of systems design.
Models are used to capture aspects of existing or envisioned systems and are simulated to predict the behavior of these systems. However, developing such models from scratch requires significant time and effort. Researchers have recognized that the time and effort can be reduced if existing models or model components are reused, leading to the study of model reusability. In this paper, we review the tasks necessary to retrieve and reuse model components from repositories, and to prepare new models and model components such that they are more amenable for future reuse.
Model reuse can be significantly enhanced by carefully characterizing the model, and capturing its meaning and intent so that potential users can determine whether the model meets their needs. Traditionally, the meaning and intent of models has been captured in textual documentation, but more recently, semantically rich, axiomatized approaches using model characteristics based on ontologies have been introduced that enable algorithmic support for the identification, discovery, and reuse of models. Researchers have also recognized that the opportunity for reuse significantly increases when models are modularized into composable model components. While many repositories and frameworks for modular modeling have been recently developed, it is important to recognize that without considerable forethought in terms of model architecture, model composability often remains elusive.
Even when models are well-characterized and organized in a repository, model users need to perform several tasks before they can successfully reuse these models and components for a specific analysis scenario. These tasks include searching, optionally adapting, composing, integrating and validating models and model components. We will review the literature on each of these tasks and focus on identifying opportunities for further improvement of model reuse support frameworks.
