DAG Based Context Reasoning: Optimised DAG Creation

Abstract:
We propose an approach to determine the pareto optimal sequences of context processing operations among distributed devices in ubiquitous computing environments with respect to multiple objective functions. The method presented is based on a graph-algorithm that processes a distributed acyclic graph that represents the communication neighbourhood among devices. Our approach is capable of respecting time-locality of distributed context providers. By this approach the quality of context is improved since the accuracy for the reasoning process increases. In our evaluation study we are able to unveil significant improvements in energy consumption and bandwidth usage.