Software and Simulation Modeling for Real-Time Software-Intensive Systems

Abstract

Successful development of large-scale complex and distributed real-time systems commonly relies on models developed separately for simulation studies and software implementation. Systems theory provides sound modeling principles to characterize structural and behavioral aspects of systems across time and space. The behavior of these models can be observed using simulation protocols that can correctly interpret time-based logical dynamics. Similarly, object-orientation theories and software architecture principles enable modeling static and dynamic behavior of systems. While models described either in system-theoretic or object-orientated languages may be used for both software design and simulation modeling, each has its own strengths and weaknesses. For example, a class of system-theoretic modeling approach called Discrete-event System Specification (DEVS) provides an appropriate basis to develop simulation models exhibiting concurrent and distributed behavior. Similarly, the Unified Modeling Language with real-time (UML-RT) constructs can be used to develop software design models that can be implemented and executed. Since software models are not suitable to be used as simulation models and simulation models may not adequately lend themselves to serve as software design blueprints, it is important to examine these approaches. We show some of the key shortcomings of these simulation and software design modeling approaches by developing some detailed specifications and implementation of a coffee machine with a focus on their treatment of logical and physical time.