Energy Management in Wireless Mobile Systems Using Dynamic Task Assignment

Wireless mobile sensing systems are hierarchical and heterogeneous in nature with components that have different energy and performance capabilities. In such systems allocation of tasks to different devices affects both performance and the entire system battery lifetime. In this paper we formulate the problem of optimal task assignment with objectives related to minimizing the total system energy consumption and maximizing the system lifetime as an Integer Linear Program (ILP). We describe a heuristic algorithm and two dynamic graph-based partitioning algorithms that are computationally efficient and that are able to adapt in real-time to changing system conditions. ILP based solutions are able to achieve optimal task assignment, but cannot be used in dynamically changing conditions due to their computationally expensive nature. We evaluate the performance of our three dynamic algorithms using Qualnet and show that they have up to 88% longer system lifetime than the ILP based solutions.