An Efficient Strategy for Model Composition in the Möbius Modeling Environment

Abstract

Möbius is well known as a modeling and evaluation environment for performance and dependability indicators. One of Möbius' key features is the modular and compositional approach to model definition and analysis. In particular, the modeler can define submodels using several formalisms and compose them to form the overall model of the system under analysis. The current algorithm for model composition in Möbius revealed performance issues when large systems are considered (such as in the modeling of realistic segments of energy or transportation infrastructures), due to the chosen data flow scheme. In this paper, a new algorithm for the same composition mechanism is proposed to improve efficiency. A case study is also developed to demonstrate the performance enhancements.