ACTA

ACTA is a tool for generating performance tests in a black-box setting. Efficient exploration of input space is the main challenge addressed by this tool. Usually, today’s complex software systems have a large input space while only a small portion of that space is responsible for performance issues. Locating that portion within a short period of time is challenging especially in cases that we have a limited test budget. ACTA relies on a generative kind of neural nets for automated and efficient exploration of the input space. This tool takes advantage of active learning to learn fast and with little test effort. More details are provided in this paper. ACTA.py is a propotype ipplementation that addresses the experiments presented in the paper. Work on the tool is in progress.