QEST 2018 has selected Dr Sadegh Soudjani’s paper for its Best Paper Award.
The International Conference on Quantitative Evaluation of Systems (QEST) is the leading forum on quantitative evaluation and verification of computer systems and networks through stochastic models and measurements.
The Best Paper Award is given every year to the best accepted paper that introduces innovative ideas leading to the advancement of the techniques for analysis of probalistic systems.
In the paper, “Approximate Time Bounded Reachability for CTMCs and CTMDPs: A Lyapunov Approach”, Dr Soudjani and his co-authors from the Max Planck Institute in Germany propose that concepts from Control Theory can be used for verification and synthesis of models traditionally studied in Computer Science.
They use the concept of Lyapunov functions – a major technique in control theory for establishing stability of dynamical systems – to provide approximate similarity relations for finite state probabilistic models.
Dr Soudjani writes, “In this paper we show that probabilistic bisimulation can be viewed as a projection matrix that relates the original probabilistic system with its bisimulation reduction. We then relax bisimulation to a quantitative notion, using a generalised projection operation between the two systems and provide bounds on their similarity.”
published on: 19 September 2018