Discrete time Markov chain families: modeling and verification of probabilistic software product lines

• Authors: Mahsa Varshosaz;Ramtin Khosravi
• Venue: Proceedings of the 17th International Software Product Line Conference co-located workshops
• Year: 2013

Software product line engineering (SPLE) enables systematic reuse in development of a family of related software systems by explicitly defining commonalities and variabilities among the individual products in the family. Nowadays, SPLE is used in a variety of complex domains such as avionics and automotive. As such domains include safety critical systems which exhibit probabilistic behavior, there is a major need for modeling and verification approaches dealing with probabilistic aspects of systems in the presence of variabilities. In this paper, we introduce a mathematical model, Discrete Time Markov Chain Family (DTMCF), which compactly represents the probabilistic behavior of all the products in the product line. We also provide a probabilistic model checking method to verify DTMCFs against Probabilistic Computation Tree Logic (PCTL) properties. This way, instead of verifying each product individually, the whole family is model checked at once, resulting in the set of products satisfying the desired property. This reduces the required cost for model checking by eliminating redundant processing caused by the commonalities among the products.