Although mobile apps have become extremely popular, their usability is limited by their reliance on battery power. To some extent, the usability of these devices has increased with research in battery designs, hardware energy efficiency, and operating system design. However, even the most carefully optimized systems can be undermined by an app that performs inefficient or unnecessary actions. Unfortunately, app developers do not have the expertise or tools to gain insight into how their apps consume battery energy and how that energy can be reduced. As a result, even a cursory examination of reviews of marketplace applications reveals many customer complaints related to energy usage, and energy related software bugs have been extensively reported on in the literature.

The goal of my group's research into app energy consumption has been to help developers understand how energy is consumed within their apps and change their apps in ways that will lead to reduced energy consumption. Within the scope of this goal, our work has focused on three inter-related areas of research. First, the development of techniques for measuring and visualizing energy consumption of apps at a fine-grained level (See eLens [hao13icse] and vLens [li13issta]). Second, the use of this measurement capability to learn about the relationship between energy and implementation structure, and the identification of energy related best practices for developers (See [li14icsme], [li15demobile] [sahin15jsep], [li14greens], and [gui15icse]). Third, the development of automated techniques that enable developers to transform their implementations and/or take advantage of the discovered energy saving practices (See [li14issta], [li14icse], [wan15icst]).

This work is supported, in part, by NSF grant 1321141 to the University of Southern California.