Dynamic power management (DPM) refers to the strategies employed at system level to reduce energy expenditure (i.e. to prolong battery life) in embedded systems. The trade-off involved in DPM techniques is between the reductions of energy consumption and latency suffered by the tasks. Such trade-offs need to be decided at runtime, making DPM an on-line problem. We formulate DPM as a hybrid automaton control problem and integrate stochastic control. The control strategy is learnt dynamically using stochastic learning hybrid automata (SLHA) with feedback learning algorithms. Simulation-based experiments show the expediency of the feedback systems in stationary environments. Further experiments reveal that SLHA attains better trade-offs than several former predictive algorithms under certain trace data.
Batteries; Embedded computer systems – Energy consumption; Energy conservation; Learning models (Stochastic processes)
Shukla, S. K.,
Stochastic learning feedback hybrid automata for dynamic power management in embedded systems.
Proceedings of the 2005 IEEE Mid-Summer Workshop on Soft Computing in Industrial Applications
Institute of Electrical and Electronics Engineers.