Computer Science Graduate Seminar: Industrial Applications of Probabilistic Model Checking

Tuesday, July 04, 2017, 3:00pm

Location: Computer Science Center, E3 Building, Room 9222

Speaker: Hao Wu, M.Sc.


Nowadays, due to the rapid growth of complexity of hardware/software systems, designing such systems to meet critical requirements (e.g., performance, energy efficiency) becomes an increasingly challenging work. Performance evaluation and optimization becomes a mandatory routine in the system design. This is usually done based on running simulations of very detailed prototypes using specific simulation tools. However the results obtained by simulation usually lack the information of performance bounds caused by the best and worst-case scenarios. In this talk, we demonstrate industrial applications of probabilistic model checking -a model-based approach- to evaluate and optimize the performance of systems during design phase and show how the results obtained give us an informative insight of the system design. The probabilistic models we use in our applications allow us to model dynamism and randomness, which are two key factors in modern system design. Our applications cover systems from several currently popular fields: embedded systems, computing servers, the next generation mobile networks and concurrent data structures. From these applications, we show that probabilistic model checking is an effective model-based approach to evaluate the performance of various types of modern systems and provides a uniform and elegant framework from modelling to performance evaluation.

The computer science lecturers invite interested people to join.