Self-Healing Test Automation Frameworks Using Reinforcement Learning for Full-Stack Test Automation
Hariprasad Sivaraman · Journal of Artificial Intelligence & Cloud Computing · 2022
In dynamic environments such as e-commerce, test automation frameworks often struggle as applications are continuously changing. In this paper a test automation framework is being proposed with self-healing ability that leverages Reinforcement Learning (RL) to automatically adapt test cases to new versions of interface, Application Programming Interfaces (APIs) and the database schema. With the integration of RL, the framework operates with minimal human intervention, makes it easier to generalize the results and lowers the costs related to maintaining the tests. Through the framework, examples inspired from common e-commerce use-cases are assessed and demonstrate the potential of RL to inject additional resilience & adaptive behavior in full-stack test automation.