Document Type

Conference Proceeding

Abstract

In this paper we show that stochastic learning automata based feedback control switching strategy can be used for dynamic power management (DPM) employed at the system level. DPM strategies are usually incorporated at the operating systems of embedded devices to exploit multiple power states available in today's ACPI compliant devices. The idea is to switch between power states depending on the device usage, and since device usage times are not deterministic, probabilistic techniques are often used to create stochastic strategies, or strategies that make decisions based on probabilities of device usage spans. Previous work (Irani et al., 2001) has shown how to approximate the probability distribution of device idle times and dynamically update them, and then use such knowledge in controlling power states. Here, we use stochastic learning automata (SLA) which interacts with the environment to update such probabilities, and then apply techniques similar to (Irani et al., 2001) to optimize power usage with minimal effect on response time for the devices.

Comments

©2003 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

UNLV article access

Share

COinS