Flexible Framework for Statistical Schedulability Analysis of Probabilistic Sporadic Tasks

The analysis of probabilistic schedulability explores all possible combinations of the probabilities of task attributes, which can easily lead to exponential computation time [24]. In this paper, we present a flexible schedulability analysis framework for periodic and sporadic tasks having probabilistic attributes where the computation time scales linearly in the size of analyzed systems. The framework is given in terms of a set of Parameterized Stopwatch Automata (PSA) models, which leads to a large degree of flexibility. Probability distributions for response time are generated using statistical model checking (UPPAAL SMC) while the overall schedulability can be checked using symbolic model checking (UPPAAL). We also define PoMD (percentage of missed deadlines) as a measure of the probabilistic schedulability of systems. To evaluate our approach, we compare the time used for computing response times and the analysis results using similar task models to that of a related analytical approach.