[IEEE]
The background application cache for smartphones maintains frequently accessed applications in memory, even if applications become inactive. The cache improves user responsiveness and reduces energy consumption for re-loading applications later. However, current cache management policies do not consider caching efficiency which may change dynamically, and thus fix the total number of cached applications regardless of reuse patterns. This study analyzes the reuse patterns of smartphone applications, and proposes a novel algorithm for adjusting the number of cached applications dynamically, based on the application usage patterns. The dynamic adjustment of the cache limit allows the system memory to be used effectively for page caches and other purposes, without sacrificing the cache hit rates.