Residential Demand Response Using Reinforcement Learning

Daniel O’Neill; Marco Levorato; Andrea Goldsmith; Urbashi Mitra

doi:10.1109/SMARTGRID.2010.5622078

Residential Demand Response Using Reinforcement Learning

Daniel O’Neill
, Marco Levorato
, Andrea Goldsmith
, Urbashi Mitra

Stony Brook University

Stanford University
University of Southern California

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

229 Scopus citations

Abstract

We present a novel energy management system for residential demand response. The algorithm, named CAES, reduces residential energy costs and smooths energy usage. CAES is an online learning application that implicitly estimates the impact of future energy prices and of consumer decisions on long term costs and schedules residential device usage. CAES models both energy prices and residential device usage as Markov, but does not assume knowledge of the structure or transition probabilities of these Markov chains. CAES learns continuously and adapts to individual consumer preferences and pricing modifications over time. In numerical simulations CAES reduced average end-user financial costs from 16% to 40% with respect to a price-unaware energy allocation.

Original language	English
Title of host publication	2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010
Publisher	IEEE Computer Society
Pages	409-414
Number of pages	6
ISBN (Print)	9781424465125
DOIs	https://doi.org/10.1109/SMARTGRID.2010.5622078
State	Published - 2010
Event	1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010 - Gaithersburg, MD, United States Duration: Oct 4 2010 → Oct 6 2010

Publication series

Name	2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010

Conference

Conference	1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010
Country/Territory	United States
City	Gaithersburg, MD
Period	10/4/10 → 10/6/10

Access to Document

10.1109/SMARTGRID.2010.5622078

Cite this

O’Neill, D., Levorato, M., Goldsmith, A., & Mitra, U. (2010). Residential Demand Response Using Reinforcement Learning. In 2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010 (pp. 409-414). (2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010). IEEE Computer Society. https://doi.org/10.1109/SMARTGRID.2010.5622078

@inproceedings{1493dee8893a488eb51d315e2141dade,

title = "Residential Demand Response Using Reinforcement Learning",

abstract = "We present a novel energy management system for residential demand response. The algorithm, named CAES, reduces residential energy costs and smooths energy usage. CAES is an online learning application that implicitly estimates the impact of future energy prices and of consumer decisions on long term costs and schedules residential device usage. CAES models both energy prices and residential device usage as Markov, but does not assume knowledge of the structure or transition probabilities of these Markov chains. CAES learns continuously and adapts to individual consumer preferences and pricing modifications over time. In numerical simulations CAES reduced average end-user financial costs from 16\% to 40\% with respect to a price-unaware energy allocation.",

author = "Daniel O{\textquoteright}Neill and Marco Levorato and Andrea Goldsmith and Urbashi Mitra",

note = "Publisher Copyright: {\textcopyright}2010 IEEE.; 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010 ; Conference date: 04-10-2010 Through 06-10-2010",

year = "2010",

doi = "10.1109/SMARTGRID.2010.5622078",

language = "English",

isbn = "9781424465125",

series = "2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010",

publisher = "IEEE Computer Society",

pages = "409--414",

booktitle = "2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010",

}

O’Neill, D, Levorato, M, Goldsmith, A & Mitra, U 2010, Residential Demand Response Using Reinforcement Learning. in 2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010. 2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010, IEEE Computer Society, pp. 409-414, 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010, Gaithersburg, MD, United States, 10/4/10. https://doi.org/10.1109/SMARTGRID.2010.5622078

Residential Demand Response Using Reinforcement Learning. / O’Neill, Daniel; Levorato, Marco; Goldsmith, Andrea et al.
2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010. IEEE Computer Society, 2010. p. 409-414 (2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Residential Demand Response Using Reinforcement Learning

AU - O’Neill, Daniel

AU - Levorato, Marco

AU - Goldsmith, Andrea

AU - Mitra, Urbashi

PY - 2010

Y1 - 2010

N2 - We present a novel energy management system for residential demand response. The algorithm, named CAES, reduces residential energy costs and smooths energy usage. CAES is an online learning application that implicitly estimates the impact of future energy prices and of consumer decisions on long term costs and schedules residential device usage. CAES models both energy prices and residential device usage as Markov, but does not assume knowledge of the structure or transition probabilities of these Markov chains. CAES learns continuously and adapts to individual consumer preferences and pricing modifications over time. In numerical simulations CAES reduced average end-user financial costs from 16% to 40% with respect to a price-unaware energy allocation.

AB - We present a novel energy management system for residential demand response. The algorithm, named CAES, reduces residential energy costs and smooths energy usage. CAES is an online learning application that implicitly estimates the impact of future energy prices and of consumer decisions on long term costs and schedules residential device usage. CAES models both energy prices and residential device usage as Markov, but does not assume knowledge of the structure or transition probabilities of these Markov chains. CAES learns continuously and adapts to individual consumer preferences and pricing modifications over time. In numerical simulations CAES reduced average end-user financial costs from 16% to 40% with respect to a price-unaware energy allocation.

UR - https://www.scopus.com/pages/publications/105021001585

U2 - 10.1109/SMARTGRID.2010.5622078

DO - 10.1109/SMARTGRID.2010.5622078

M3 - Conference contribution

SN - 9781424465125

T3 - 2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010

SP - 409

EP - 414

BT - 2010 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010

PB - IEEE Computer Society

T2 - 1st IEEE International Conference on Smart Grid Communications, SmartGridComm 2010

Y2 - 4 October 2010 through 6 October 2010

ER -

Residential Demand Response Using Reinforcement Learning

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this