A Rejuvenation Strategy in Android

Compared to critical systems, the user of Android is more concerned about the application launch time. It needs to perform rejuvenation to improve the user experience when the Android suffers from software aging that would lead to an increase of the launch time. Also, rejuvenation can incur in some costs. So it is important to make an optimal rejuvenation strategy to minimize these costs and improve the user experience. At the same time, different workloads on Android have different aging phenomena, which may influence the validation of system state. Based on these, we carried out experiments to simulate user behavior on Android and find software aging phenomena in Android. Then we used active learning based on random forest and four states to build model and constructed software aging rejuvenation strategy on Android. Comparing the experiments without rejuvenation, there is an obvious decrease of the average of application launch time after using the rejuvenation strategy, and the running time of experiments with rejuvenation is longer than that without rejuvenation, which improves the user experience and reduces the probability of unexpected errors.