Skip to Main content Skip to Navigation

A Discrete Time Markov Decision Process for Energy Minimization Under Deadline Constraints

Bruno Gaujal 1 Alain Girault 2 Stéphan Plassart 1, 2
1 POLARIS - Performance analysis and optimization of LARge Infrastructures and Systems
LIG - Laboratoire d'Informatique de Grenoble, Inria Grenoble - Rhône-Alpes
2 SPADES - Sound Programming of Adaptive Dependable Embedded Systems
Inria Grenoble - Rhône-Alpes, LIG - Laboratoire d'Informatique de Grenoble
Abstract : This paper proposes a Discrete Time Markov Decision Process (MDP) approach to compute the optimal on-line speed scaling policy to minimize the energy consumption of a single processor executing a finite or infinite set of jobs with real-time constraints. We provide several qualitative properties of the optimal policy: monotonicity with respect to the jobs parameters, comparison with on-line deterministic algorithms. Numerical experiments in several scenarios show that our proposition performs well when compared with off-line optimal solutions and out-performs on-line solutions oblivious to statistical information on the jobs.
Complete list of metadatas

Cited literature [29 references]  Display  Hide  Download
Contributor : Stéphan Plassart <>
Submitted on : Tuesday, December 3, 2019 - 5:40:46 PM
Last modification on : Monday, April 20, 2020 - 11:44:05 AM
Document(s) archivé(s) le : Wednesday, March 4, 2020 - 5:56:11 PM


Files produced by the author(s)


  • HAL Id : hal-02391948, version 1


Bruno Gaujal, Alain Girault, Stéphan Plassart. A Discrete Time Markov Decision Process for Energy Minimization Under Deadline Constraints. [Research Report] RR-9309, Grenoble Alpes; Inria Grenoble Rhône-Alpes, Université de Grenoble. 2019, pp.46. ⟨hal-02391948⟩



Record views


Files downloads