Adaptive Online Power Management Algorithms

S. Kim (USA)


Real time system, Dynamic voltage processor, On-line decisions, Adaptive task scheduling, Load balancing


Design and analysis of power management techniques is an important and widely open research field. Energy consumption, in particular, has become an increasingly important consideration in many real-time systems. In this paper, I propose new on-line power management algorithms that manage heterogeneous tasks and adapt their execution to maximize energy efficiency. In addition, I consider other performance criteria such as delay overhead. The most important feature of my proposed scheme is that various online management techniques - processor capacity management, request admission, task scheduling and delay control mechanism - are developed and integrated for higher system performance and energy efficiency. Simulation results clearly indicate the superior performance of my proposed scheme to strike the appropriate performance balance bewteen contradictory requiremenst.

Important Links:

Go Back