QExplore: An exploration strategy for dynamic web applications using guided search.

Dynamic exploration approaches play an important role in automating web testing and analysis. They are extensively used to explore the state-space of a web application for achieving complete coverage of the application's functionality. Dynamic exploration approaches support end-to-end automation of testing to verify the correct behavior of a web application. However, existing approaches failed to explore the states behind the web forms and can get stuck in dynamic regions of web applications resulting in poor functionality coverage and diversity. Consequently, existing approaches are regressive in nature and sensitive to small DOM mutations which may not be interesting from a testing perspective. In this paper, we propose a dynamic exploration approach using guided search inspired by Q-learning that systematically explores dynamic web applications requiring less or no prior knowledge about the application. Our approach is implemented in a tool called QExplore and is empirically evaluated with six popular open-source and one real industrial application. The results show that QExplore achieved higher coverage with more diverse DOM than the existing state-of-the-art tools Crawljax and WebExplor. QExplore also results in a greater number of navigational paths, error states and distinct DOM states when compared with the existing tools.(c) 2022 Elsevier Inc. All rights reserved.