Power saving for multiple MEMS-based devices by probabilistic data partition

This paper considers an optimization problem for minimizing the power consumption of MEMS-based devices. We first map the time-consuming problem from the integer space Zⁿ to the Euclidean space Rⁿ and employ a gradient-based method to solve the mapped problem in Rⁿ. The great strength of the proposed method lies in its linear convergence. Experimental results show that significant energy can be saved by shutting down as many unnecessary MEMS-based devices as possible. The results also suggest that other similar optimization problems can be solved by the same way.