Static/Dynamic Analysis of Android Applications to Improve Energy-Efficiency

Abstract

Energy consumption of mobile applications has a great impact on the satisfaction of users. Many developers do not consider energy-efficient programming practices when coding an applications. Therefore, many applications can be improved in terms of energy usage by detecting and removing their energy defects. Most works towards this goal focus on static analysis of Android applications. In this paper, we propose a hybrid framework consisting of static and dynamic modules for finding energy defects. Our static module can effectively analyse an Android application considering its life-cycle and look for the energy anti-patterns. The results of static analysis are then transformed to executable GUI-based test cases. In the dynamic module, the generated test cases are run, and their results are evaluated using an automated oracle, which does not need the energy measurement. Using our framework, we have analyzed 29 applications, and found 9 applications with energy defects. Our dynamic module could effectively detect the false positives in the results of the static analysis module.