A Survey on Energy Leak Detection in Android Applications: Approaches and Challenges

Abstract

With the rapid growth in mobile application usage, energy efficiency in Android applications has become a critical research focus. Excessive battery consumption, often due to energy leaks such as mismanaged wake-locks and other resource inefficiencies, degrades user experience and device longevity. This survey reviews current methodologies for detecting energy leaks in Android applications, including static analysis, dynamic analysis, hybrid analysis, machine learning approaches, and testing frameworks for identifying energy-related code smells. Static analysis examines code without execution to detect potential energy inefficiencies, while dynamic analysis observes app behavior during runtime to identify actual energy drains. Hybrid analysis combines both approaches, enhancing detection accuracy. Recently, machine learning models have been applied to analyze app performance data, shifting the focus from traditional testing to predictive diagnostics. This survey highlights key advancements, challenges, and emerging trends in energy leak detection and advocates for integrating machine learning algorithms. By leveraging app usage data, machine learning offers a scalable, accurate, and proactive solution to energy inefficiency in Android applications, paving the way for more sustainable mobile development.