Dynamic Power Management is a design methodology aiming at reducing power consumption of electronic systems, by performing selective shutdown of the idle system resources. The effectiveness of a power-management scheme depends critically on an accurate modeling of the environment, and on the computation of the control policy. This paper presents two methods for characterizing non-stationary service requests by means of a prediction scheme based on sliding windows. Moreover, it describes how control policies for nonstationary models can be derived.
|Title of host publication||Design, Automation, and Test in Europe|
|Subtitle of host publication||The Most Influential Papers of 10 Years Date|
|Number of pages||12|
|Publication status||Published - 2008 Dec 1|
All Science Journal Classification (ASJC) codes