ra4xstate: An Efficient Quantitative Robustness Analysis Approach for Statecharts

Robustness analysis is a part of the validation process that includes testing the behavior of a system against its specification under unexpected conditions in order to check whether the system fulfills robustness requirements or not. This paper proposes ra4xstate, a robustness analysis framework in the context of Model-Driven Development, for finite state machines. Our approach takes a behavioural and a property model of the system under the test as inputs and evaluates the robustness of the system based on a notion of the cost that is computed for every off-track execution step. The experimental results show that compared to the traditional approaches that annotate traces with timestamps and variable values ra4xstate detects almost all nonrobust instances while reducing the size of the trace significantly and incurring similar runtime overhead.