Periodic Reporting for period 1 - AST (Automatic System Testing)
Reporting period: 2019-01-01 to 2020-06-30
The AST project addresses the development of an automatic GUI testing solution that uses Q-learning and semantic information to generate test cases for Web applications. Q-learning is exploited to incrementally and continuously learn how an application under test reacts to the generated test inputs and thus incrementally generate more effective and more sophisticated combinations of inputs. Semantic information, such as business rules and databases of semantically relevant input values, is used to guide the testing activity towards meaningful interaction sequences, limiting the generation of irrelevant inputs and interactions. The resulting technology has been experimented with industrial partners producing initial evidence of its suitability for the industrial context. Indeed, the GUI test automation solution investigated in the AST project can alleviate testers from part of their testing effort and can help rapidly releasing high-quality Web applications. In addition to its direct impact on software developers, the project indirectly affects the quality of the experience of the users of Web applications who can benefit from the availability of more reliable software applications.