2019/08/19 by Mahshid Helali Moghadam, Moghadam, Mahshid Helali, Mehrdad Saadatmand +7 · 1 citation
Computer Science · #Software System Performance and Reliability #Software Testing and Debugging Techniques #Software Engineering Research
paper · pdf · doi:10.48550/arxiv.1908.06900
Test automation brings the potential to reduce costs and human effort, but\nseveral aspects of software testing remain challenging to automate. One such\nexample is automated performance testing to find performance breaking points.\nCurrent approaches to tackle automated generation of performance test cases\nmainly involve using source code or system model analysis or use-case based\ntechniques. However, source code and system models might not always be\navailable at testing time. On the other hand, if the optimal performance\ntesting policy for the intended objective in a testing process instead could be\nlearned by the testing system, then test automation without advanced\nperformance models could be possible. Furthermore, the learned policy could\nlater be reused for similar software systems under test, thus leading to higher\ntest efficiency. We propose SaFReL, a self-adaptive fuzzy reinforcement\nlearning-based performance testing framework. SaFReL learns the optimal policy\nto generate performance test cases through an initial learning phase, then\nreuses it during a transfer learning phase, while keeping the learning running\nand updating the policy in the long term. Through multiple experiments on a\nsimulated environment, we demonstrate that our approach generates the target\nperformance test cases for different programs more efficiently than a typical\ntesting process, and performs adaptively without access to source code and\nperformance models.\n