Optimizing the Symbolic Execution of Evolving Rhapsody Statecharts

Thumbnail Image
Khalil, Amal
Symbolic Execution , Rhapsody Statecharts , Model-based Analysis , Incremental Analysis , Model-driven Engineering
Model Driven Engineering (MDE) is an iterative and incremental software development process. Supporting the analysis and the verification of software systems developed following the MDE paradigm requires to adopt incrementality when carrying out these crucial tasks in a more optimized way. Communicating State Machines are one of the various formalisms used in MDE tools to model and describe the behavior of distributed, concurrent and real-time reactive systems (e.g., automotive and avionics systems). Modeling the overall behavior of such systems is carried out in a modular way and on different levels of abstraction (i.e., it starts with modeling the behavior of the individual objects in the system first then modeling the interaction between these objects). Similarly, analyzing and verifying the correctness of the developed models to ensure their quality and their integrity is performed on two main levels. The intra-level is used to analyze the correctness of the individual models in isolation of the others, while the inter-level is used to analyze the overall interoperability of those that are communicating with each other. One way to facilitate the analysis of the overall behavior of a system of communicating state machines is to build the global state space (also known as the global reachability tree) of the system. This process is very expensive and in some cases it may suffer from the state explosion problem. Symbolic execution is a technique that can be used to construct an abstract and a bounded version of the system global state space that is known as a symbolic execution tree (SET), yet the size of the generated trees can be very large especially with big and complex systems that are composed of multiple objects. As the system evolves, one way to avoid regenerating the entire SET and repeating any SET-based analyses that have been already conducted is to utilize the previous SET and its analysis results in optimizing the process of generating the SET of the system after the change. In this thesis, we propose two optimization techniques to direct the successive runs of the symbolic execution technique towards the impacted parts of an evolving state machine model using memoization (MSE) and dependency analysis (DSE), respectively. The evaluation results of both techniques showed significant reduction in some cases compared with the standard symbolic execution technique.
External DOI