Mobile Learning Support with Statistical Inference-Based Cache Management

摘要: Supporting efficient data access in the mobile learning environment is becoming a hot research problem recent years, and becomes tougher when clients are light-weight devices such as cell phones whose limited storage space prevents from holding large cache. A practical solution to store cache at some proxies nearby, so that can these instead of servers order reduce latency time. However, move freely, may not enhance overall performance because it become too far away for access. In this paper, we propose statistical caching mechanism which makes use prior knowledge (statistical data) predict pattern user movement then replicates/migrates objects among different proxies. We inference based heuristic search algorithm accommodate dynamic environment. Experimental studies show that, with an acceptable complexity, our obtain good on data.