Improving Energy Consumption in Android Apps

Mobile applications sometimes exhibit behaviors that can be attributed to energy bugs depending on developer implementation decisions. In other words, certain design decisions that are technically "correct" might affect the energy performance of applications. Such choices include selection of color palettes, libraries used, API usage and task scheduling order. We study the energy consumption of Android apps using a power model based on a multi-objective approach that minimizes the energy consumption, maximizes the contrast, and minimizes the distance between the chosen colors by comparing the new options to the original palette. In addition, the usage of unnecessary resources can also be a cause of energy bugs depending on whether or not these are implemented correctly. We present an opportunity for continuous investigation of energy bugs by analyzing components in the background during execution on Android applications. This includes a potential new taxonomy type that is not covered by state-of-the-art approaches.