Agent Driven Scenario Testing

The scripting of automated test scenarios is a common practice in nowadays software development patterns. These aid the validation of specific use patters that a user can be expected to, or is able to, perform. However, for complex systems it is safe to assume that the definition of realistic use patters is not only a tedious and time consuming goal to chance, but also one that is unlikely to ever be reached.

Hence, this Master Project aims at developing a testing infrastructure, in which intelligent agents are, first derived for the behavioural patterns produced by real users whilst using a target software, and later utilised to emulate realistic user decision makers in a simulation environment.

The project would require students to:

• Develop a reusable strategy for the weaving of user analytics recording logic in the codebase of a selected target software.
• Synthesise realistic agents from the behavioural examples collected from real world users
• Develop a user emulator, capable of interacting with the selected target software, and sample action from an agent process
• Package agents and user emulators in a simulation environment in which the target software can be tested

This project will be supervised by Prof. Dr. Harald C. Gall, Marco Edoardo Palma, and Alex Wolf.

Please reach out to Marco Edoardo Palma or Dr. Pasquale Salza if you have further questions about the project.