Probabilistic Analysis of Time-Sensitive Systems

We will explore predictive techniques to verify that these systems satisfy their requirements. Our verification techniques will provide greater confidence in the achievement of mission goals by complex distributed stochastic systems fielded by the U.S. Department of Defense (DoD). For large-scale data analysis applications, our techniques will make it possible to cost-effectively provision systems on “right-sized” computational platforms, providing predictable performance at optimum cost. We will pursue two tasks: (A): for real-time safety-critical systems, we will use statistical model checking and uncertainty quantification to predict their behavior; (B): for data-intensive applications, throughput-oriented techniques will be used.