Intelligent Energy Management
Submitting InstitutionUniversity of Southampton
Unit of AssessmentComputer Science and Informatics
Summary Impact TypeSocietal
Research Subject Area(s)
Information and Computing Sciences: Artificial Intelligence and Image Processing, Information Systems
Engineering: Electrical and Electronic Engineering
Summary of the impact
Research at the University of Southampton, into the engineering of
complex socio-technical systems, has underpinned new technologies in the
area of intelligent energy management, and made Professors Nick Jennings
and Alex Rogers trusted sources of advice for energy policymakers, key
stakeholders and industrial researchers. The work has had an economic,
environmental and societal impact: it has shaped R&D strategies of
leading British companies like BAE Systems and Secure Meters; the launch
of iPhone apps and websites have supplied private and industrial users
with personalised data regarding their energy use, resulting in cost
savings and reductions in carbon emissions; it has enabled charities to
provide energy-saving advice to households directly; and has won an
international technology showcase competition leading to a spinout and
commercialisation of research.
Department of Energy and Climate Change findings showed UK emissions of
greenhouse gases covered by the Kyoto Protocol rose by 3.5 % between 2011
and 2012 - primarily resulting from greater use of coal for electricity
generation at power stations and an increase in residential gas use.
Research conducted within the 'energy' domain of the University of
Southampton's Agents, Interaction and Complexity Group (AICG) has focused
on the development of key tools for energy management that will allow
users to defer use of electricity to off-peak times, to use energy more
effectively and to reduce their carbon footprint. The programme was led by
Professors Nick Jennings, Professor of Computer Science (since 1999), and
Alex Rogers, Professor of Computer Science (since 2013, Lecturer since
The starting point for this research was the development of key
technologies as part of two projects led by Jennings in conjunction with
the University of Oxford: ARGUS II (2003-2008, A) and ALADDIN (2005-2011,
B). This work was concerned with the development of software agents that
act on the behalf of humans with minimal intervention. Researchers
addressed fundamental questions regarding the engineering of systems
composed of multiple autonomous agents and how individual agents should
make effective decisions in the face of uncertain and dynamic
environments. Key findings included the development of software that uses
local message-passing algorithms to allow agents to communicate with each
other, to solve complex problems involving uncertainty [3.1] and
the use of Gaussian processes to extend the use of independent agents, and
uncertainty-based reasoning and prediction, into areas where resources are
limited and continually shifting [3.2].
Three further interconnected AICG projects have focussed on the specific
requirements of `intelligent energy management': `Intelligent
Decentralised Energy-Aware Systems' (iDEaS) project (2009-ongoing, C),
`Intelligent Agents for Home Energy' (IAHEM) (2010-2013, D) and ORCHID
(2011-ongoing, E), in which researchers built and applied agent approaches
to real-world applications in the critical domains of the smart grid and
citizen science in conjunction with the Universities of Oxford and
Through the development of novel machine-learning algorithms, using
Gaussian processes to estimate and predict energy use, both in the form of
electricity-demand within the grid and the thermal characteristics of
building and homes, researchers designed a tool through which heating and
electrical loads can be optimised and deferred. This allows users to
reduce energy and carbon emissions without impacting on comfort [3.3,
3.6]. Researchers also developed novel optimisation routines that
allow users to make optimal use of energy storage devices, by enabling
cheap and low-carbon electricity to be stored and loads to be deferred,
resulting in minimised carbon emissions [3.4]. Researchers'
application of techniques from game theory and mechanism design resulted
in the development of a range of novel online auction mechanisms that can
automatically allocate and price electric vehicle charging. This ensures
that individual users get the best deal possible while satisfying the
constraints of the distribution network [3.5].
References to the research
(best three are starred)
3.1 *A. Rogers, A. Farinelli, R. Stranders and N. R. Jennings (2011) "Bounded
approximate decentralised coordination via the max-sum algorithm" Artificial
Intelligence 175 (2) 730-759.
3.2 M. A. Osborne, S. J. Roberts, A. Rogers and N. R. Jennings (2012) "Real-time
information processing of environmental sensor network data using
Bayesian Gaussian processes" ACM Trans on Sensor Networks 9
3.3 S. Ramchurn, P. Vytelingum, A. Rogers and N. R. Jennings (2011) Agent-based homeostatic
control for green energy in the smart grid. ACM Transactions on
Intelligent Systems and Technology, 2 (4).
3.4 *P. Vytelingum, T. Voice, S. Ramchurn, A. Rogers and N. R. Jennings
(2011) Theoretical and Practical Foundations of Agent-Based
Micro-Storage in the Smart Grid. Journal of Artificial Intelligence
Research, 42, 765-813.
3.5 *V. Robu, E. H. Gerding, S. Stein, D. C. Parkes, A. Rogers and N. R.
Jennings (2013) An Online Mechanism for Multi-Unit Demand and its
Application to Plug-in Hybrid Electric Vehicle Charging. Journal
of Artificial Intelligence Research (In Press).
3.6 Ramchurn, S., Vytelingum, P., Rogers, A. and Jennings, N. (2011) Putting
the "Smarts" into the Smart Grid: A Grand Challenge for Artificial
Intelligence. Communications of the ACM, 55 (4), 86-97.
A. ARGUS: PI Jennings; "Decentralised Data Fusion"; EPSRC/MoD/DTI
(GR/S20727/01); 2003-2008; £1m.
B. ALADDIN: PI Jennings; CoIs: Rogers, Ramchurn, Polukarov;
"Autonomous Learning Agents for Decentralised Data and Information
Systems"; BAE SYSTEMS/EPSRC (EP/C548051/1);
C. iDEaS: PI Jennings & Rogers; "Intelligent Decentralised
Energy-Aware System"; Industrially funded by Secure Meters; 2009-2014;
D. IAHEM: PI Rogers; CoI Jennings; "Intelligent Agents for Home
Energy Management"; EPSRC (EP/I000143/1); 2010-2013; £0.8m.
E. ORCHID: PI Jennings; CoIs: Rogers, Ramchurn; "Human-Agent
Collectives"; EPSRC Programme Grant (EP/I011587/1); 2011-2016; £5.5m.
Details of the impact
AICG research has resulted in a range of algorithms and technological
strategies that allow individual energy consumers and suppliers to
coordinate their actions within electricity networks and electricity
markets to make better use of limited or constrained resources. This has
led to economic savings in energy expenditure and has contributed to the
reduction of carbon emissions through work on "smart meters", which the
Government aims to have in 26 million homes by 2020.
An early key output of this work was GridCarbon: an iPhone
application for monitoring the UK electricity grid's carbon intensity and
generation mix, launched in January 2010. GridCarbon was the first
smartphone app aimed at electricity distribution professionals, to provide
real-time information on a national electricity grid. The last update in
October 2013 currently has over 2,000 users [5.1]. In July 2013, GridCarbon
was cited during the House of Lords' Grand Committee on the Energy Bill,
in the context of the need for the Government to develop a gold-standard
methodology for providing real time grid information to users [5.2].
During 2012/13, live energy displays using data derived from GridCarbon
were used during a six-month trial at a Federal Mogul (automotive and
aerospace product supplier) factory in Chapel-en-le-Frith. During the
trial, live energy usage data was combined with live production data to
generate per component energy feedback to factory workers, as part of a
company wide initiative to pinpoint where energy was being wasted. The
trial reduced energy waste by up to 20%, and in June 2013 the approach was
rolled out across the whole factory, with a view to extending it to other
Federal Mogul sites. [5.3, 5.8]
In December 2012, AICG launched MyJoulo, a free online energy
advice system. Drawing on AICG research, MyJoulo uses intelligent
algorithms to analyse data collected from an AICG-designed USB
temperature logger, to build a thermal model of the home and to infer the
operational settings of the heating system. This model is then used to
calculate the impact, in terms of percentage reduction in heating costs,
of various interventions such as adjusting timer settings. In its first
four months of operation, MyJoulo provided advice to over 750
private households, identifying over £50k in annual heating bill savings.
Feedback from public users includes: "Geekery at its best." [5.4].
In September 2013, MyJoulo won first place in the British Gas
Connecting Homes competition, and was named as the UK's brightest home
tech startup, at an event which saw 25 companies from around the world
pitch innovative products and services in the home energy sector to
British Gas. MyJoulo won £30k in seed funding at the event, was
spun-out of University of Southampton in November 2013, and plans are
progressing for a trial with British Gas customers in January 2014 [5.5].
In June 2011, Rogers participated in an industry workshop organised by
the UK Energy Research Centre, to discuss how smart meter research can
reduce customers' energy demands [5.6]. Subsequently, in March
2013, AICG began work with the Centre for Sustainable Energy, a charity
that provides energy advice to individuals and organisations, to extend
intelligent energy applications to electricity monitoring, using a bespoke
electricity logger developed at AICG [5.9]. In March 2013, AICG
was also invited by the Department of Energy and Climate Change to
participate in the next stage of its 'Smart Heating Control Trial'
providing sets of three MyJoulo loggers to 21 households, and
providing customised energy and heating system use analysis to the
customer experience team, in the first stage of the trial [5.10].
Throughout the course of this research, AICG researchers have
collaborated with industrial partners, influencing their funding decisions
and shaping their business strategies in the intelligent energy sector.
Having secured a £1.6 million grant from Secure Meters (market leader in
supply of intelligent electronic meters) in 2009, they worked with the
company's engineers to use Gaussian process prediction algorithms to
predict aggregate electricity demand in the UK and Indian electricity
grid. Utilita, an energy retailer and Secure Meters subsidiary, applied
AICG clustering algorithms to the problem of clustering UK consumers
according to their electricity and gas consumption. Secure Meters
engineers incorporated both novel algorithms into their standard toolset [5.11].
AICG has worked alongside developers from Horstmann Controls (the UK's
leading home heating controller manufacturer and a wholly-owned subsidiary
of Secure Meters) to demonstrate the application of intelligent learning
algorithms within Horstmann's heating controllers. Horstmann says the
on-going trials will inform the design of its next-generation home heating
controllers [5.12].The Southampton researchers are also working
with BAE Systems within an EPSRC Impact Acceleration Account-funded
knowledge transfer secondment, to deploy advanced load recognition
algorithms within their DEMS (Deployed Energy Management System), and with
Microsoft Research to deploy algorithms to learn the thermal performance
of homes within their HomeLab smart home infrastructure.
AICG researchers have proactively disseminated their research findings to
increase awareness among the general public and industry of the value of
smart meters. The research has appeared in BBC News, The Economist,
New Scientist, The Engineer (monthly e-magazine, 138,000
subscribers) and Science Daily (3 million monthly visitors
online). Alex Rogers wrote about MyJoulo in the Huffington
Post's Tech section (28/3/13). Academics delivered presentations at
Autonomous Agents and Multi-Agent Systems (AAMAS) conferences in 2010,
2011 and in 2012 (where Rogers gave the keynote address 'Delivering the
Smart Grid'). Each conference attracted several hundred leading
researchers and practitioners of agent technology [5.7].
Sources to corroborate the impact
[5.1] GridCarbon. http://itunes.apple.com/gb/app/gridcarbon/id346832866/.
GridCarbon is cited by Lord Grantchester in Column GC398
[5.3] Building Banter.
[5.4] MyJoulo. http://www.myjoulo.com.
[5.5] MyJoulo is named the UK's brightest home tech startup.
[5.7] Public and industry engagement:
Huffington Post (28/03/13) http://www.huffingtonpost.co.uk/dr-alex-rogers/artificial-intelligence-ai_b_2969797.html
BBC News 8/12/12 http://www.bbc.co.uk/news/technology-20173641
The Economist 06/09/11: http://www.economist.com/blogs/babbage/2011/09/recharging-electric-cars
The Engineer 06/05/2011: http://www.theengineer.co.uk/sectors/automotive/news/computerised-scheduling-for-electric-vehicle-charging/1008558.article
New Scientist 07/03/2011: http://www.newscientist.com/article/mg20928025.700-smartgrid-stockbrokers-to-manage-your-power.html
Science Daily 05/06/2010: http://www.sciencedaily.com/releases/2010/05/100505092527.htm
AAMAS 2012 conference:
[5.8] Chief Technical Officer, Moixa Technology Ltd.
[5.9] Senior Development Manager, Centre for Sustainable Energy
[5.10] Head of Consumer Insight, Energy Technologies Institute
[5.11] Managing Director, Secure Meters
[5.12] Product & Market Development Director, Horstmann Controls.