INTRODUCTIONRESEARCHRECENT PROJECTSCOURSESPUBLICATIONS
Smart Devices Optimization
HOME : RESEARCH : Smart Devices Optimization
 Personalized Optimization for Smart Devices
1. Motivation
  • Smartphones are truly personalized computing devices
  • Only one dominant user per smartphone
  • As a highly personalized computing device, smartphones present a unique new opportunity for system optimization
  • In our own study on the Android app usage profiling of 30 participants,
  • A single user exhibits very regular usage patterns
  • Only a small number of different apps are used
  • A small set of distinctive app usage patterns is repeatedly appeared
  • Some apps are strongly related to each other
  • Different users are quite different in their usage patterns
  • 2. Research goal
  • Collecting and identifying distinctive usage pattern during runtime for personalized optimization
  • 3. POA framework
  • Personalized Optimization framework for Android Smartphones (POA)

  • 4. App usage modeling construction
  • App Usage Model Module
  • Accumulating past app sequence
  • Dynamically building a user-specific AUM based on the past app sequence
  • An example of building an app usage model
  • 5. Usage model-based optimization
  • Consisting of various optimization-specific sub-modules for personalized optimization
  • When each sub-module applies the user-specific AUM , each sub-module requests hints on future app executions from the app usage modeling module
  • 6. Case study: Optimizing restart ratio of apps.
  • The Android default app killer
  • Limiting the number of the backgrounds apps lower than the predefined maximum number (15)
  • Terminating an app which is least recently used (LRU)
  • The problem of the default app killer
  • Restart ratio of apps is high because it does not recognize the single user's app usage tendency (In our study, most users experience restart ratio more than 10%)
  • Optimizing the restart ratio exploiting our proposed AUM
  • Improving the performance of Android's app killing/restarting policy
  • 7. International contest
  • Wook Song, Yeseong Kim, Hakbong Kim, Jehun Lim, Sang-Woo Jun, and Jihong Kim, "Personalized Optimization for Android Smartphones," Android Competition in Embedded Systems Week 2011, Taipei, Taiwan, October 12, 2011

  •