A method and device for automated testing

Abstract

The invention discloses a method and device for automatic testing, and relates to the technical field of computers. A specific implementation of the method includes: collecting decision-making trajectory data of testers during testing, and using the decision-making trajectory data as an expert strategy; creating a model of a reward function; using the expert strategy as an input of an inverse reinforcement learning algorithm, To find the reward coefficient that maximizes the value of the reward function; according to the reward function corresponding to the maximum reward coefficient, calculate the strategy corresponding to the maximum cumulative reward function value, so as to perform automated testing according to the strategy. In this embodiment, a reward function is designed through an inverse reinforcement learning algorithm, and then a strategy corresponding to the expert strategy is obtained, and automated testing is performed according to the strategy. The human test results were similar.

Description

technical field [0001] The invention relates to the field of computers, in particular to a method and device for automatic testing. Background technique [0002] In software companies, whether it is the release of new functions or the return of old functions, a large number of tests are required to test the stability and robustness of the software. When conducting software testing, the more and deeper the test scenarios entered and the wider the coverage, the better the test effect can be considered. [0003] There are three ways in the prior art when performing software testing. The first is manual testing: the tester clicks each button to reach each scene; the second is testing with automated tools: such as Android's Monkey, which simulates the user's key input, touch screen input, gesture input, etc., depending on the number of devices There will be abnormalities for a long time; the third is to use the reinforcement learning algorithm, set the reward function of the al...

AI Technical Summary

Problems solved by technology

[0004] In the process of realizing the present invention, the inventor found that there are at least the following problems in the prior art: the first method is costly and will take up a lot of testing resources

The second method is highly random, and the test depth and coverage cannot be considered at the same time; the third method, it is difficult to quantify what kind of action is good and what kind of action is bad in some test scenarios, and it is difficult to set Reasonable Reward Function

Images