Robustness of Model-Based Simulations

—This paper proposes a framework for determining the correctness and robustness of simulations of hybrid systems. The focus is on simulations generated from model-based design environments and, in particular, Simulink. The correctness and robustness of the simulation is guaranteed against floatingpoint rounding errors and system modeling uncertainties. Toward that goal, self-validated arithmetics, such as interval and affine arithmetic, are employed for guaranteed simulation of discrete-time hybrid systems. In the case of continuous-time hybrid systems, self-validated arithmetics are utilized for overapproximations of reachability computations. Keywords-Robustness; Simulation; Hybrid systems; Model Validation and Analysis; Floating-point arithmetic;

• Type: Conference
• Year: 2009
• Where: RTSS
• Authors: Georgios E. Fainekos, Sriram Sankaranarayanan, Franjo Ivancic, Aarti Gupta
• Keywords: Applied Computing, Hybrid Systems, RTSS 2009, Self-validated Arithmetics, Simulation