With the popularity of deep reinforcement learning(DRL), people have great interest in using deep reinforcement learning for application automated testing. However, most automated testing methods based on reinforcement learning ignore text information, use random sampling in experience replay and ignore the characteristics of Android automated testing. To solve above problem, this paper proposes ITPRTesting(Integrated Text feature information and Priority experience in Testing). It extracts the text information in the interface and uses the BERT algorithm to generate sentence vectors. It fuses the interactive control feature diagram(ICFD), which is mentioned in the previous work, and text information as the state required by reinforcement learning. And in reinforcement learning, the priority experience replay is combined, also the traditional priority experience replay is improved. This paper has carried out experiments on 10 open source applications. The experimental results show that ITPRTesting is superior to other methods in statement coverage and branch coverage.

With the improvements of Deep Reinforcement Learning (DRL), there have been tremendous interests in utilizing DRL for automated application testing. However, most automated testing methods based on reinforcement learning have the problem of interacting with invalid UI areas and invalid interactions with controls. To solve this problem, this paper extracts the page features, constructs the Interactive Control Feature Diagram(ICCD); improves the DDQN network structure, adds the residual network, makes the algorithm take the picture as the input, and splits the original single output action(n*w*h) into two successive outputs: the interaction(1,n) and the position(1,w*h); a new reward function which combines the interaction times and the image similarity of ICCD is proposed to explore different UIs and ensure that there will be more than one action will be executed under the same UI. Experiments are carried out on five open source applications. The experimental results show that the proposed method is superior to other methods in code coverage and branch coverage.