MARTE Genetic Algorithm for Uncovering Scenarios Leading to Data Races in Concurrent Systems

Identifying the concurrency problems early in the design process is most important. Because, these problems caused to disturbance or damage when building larger and more complex systems. The recent trend is Model Driven Development (MDD), so there is a method used for the detecting concurrency problems, which is based on design models articulated in Unified Modelling Language (UML). The UML notation is not enough to model a system for s given purpose, so the notation is extended via profiles. The system aim is to develop a scalable integrated method to tailor all concurrency problems. This is achieved through three steps. First one is extracting all relevant concurrency information using UML/Modelling and Analysis of RealTime and Embedded Systems (MARTE) model diagrams. Second one is detecting all concurrency faults by using the search-based technique. Finally demonstrating scalability in terms of fault detection. Apart from Dead lock and Starvation, also detecting one more concurrency problem such as Data Races.