The University of Southampton
University of Southampton Institutional Repository

Mechanism design for information aggregation within the smart grid

Mechanism design for information aggregation within the smart grid
Mechanism design for information aggregation within the smart grid
The introduction of a smart electricity grid enables a greater amount of information exchange between consumers and their suppliers. This can be exploited by novel aggregation services to save money by more optimally purchasing electricity for those consumers. Now, if the aggregation service pays consumers for said information, then both parties could benefit. However, any such payment mechanism must be carefully designed to encourage the customers (say, home-owners) to exert effort in gathering information and then to truthfully report it to the aggregator. This work develops a model of the information aggregation problem where each home has an autonomous home agent, which acts on its behalf to gather information and report it to the aggregation agent. The aggregator has its own historical consumption information for each house under its service, so it can make an imprecise estimate of the future aggregate consumption of the houses for which it is responsible. However, it uses the information sent by the home agents in order to make a more precise estimate and, in return, gives each home agent a reward whose amount is determined by the payment mechanism in use by the aggregator. There are three desirable properties of a mechanism that this work considers: budget balance (the aggregator does not reward the agents more than it saves), incentive compatibility (agents are encouraged to report truthfully), and finally individual rationality (the payments to the home agents must outweigh their incurred costs). In this thesis, mechanism design is used to develop and analyse two mechanisms. The first, named the uniform mechanism, divides the savings made by the aggregator equally among the houses. This is both Nash incentive compatible, strongly budget balanced and individually rational. However, the agents' rewards are not fair insofar as each agent is rewarded equally irrespective of that agent's actual contribution to the savings. This results in a smaller incentive for agents to produce precise reports. Moreover, it encourages undesirable behaviour from agents who are able to make the loads placed upon the grid more volatile such that they are harder to predict. To resolve these issues, a novel scoring rule-based mechanism named sum of others' plus max is developed, which uses the spherical scoring rule to more fairly distribute rewards to agents based on the accuracy and precision of their individual reports. This mechanism is weakly budget balanced, dominant strategy incentive compatible and individually rational. Moreover, it encourages agents to make their loads less volatile, such that they are more predictable. This has obvious advantages to the electricity grid. For example, the amount of spinning reserve generation can be reduced, reducing the carbon output of the grid and the cost per unit of electricity. This work makes use of both theoretical and empirical analysis in order to evaluate the aforementioned mechanisms. Theoretical analysis is used in order to prove budget balance, individual rationality and incentive compatibility. However, theoretical evaluation of the equilibrium strategies within each of the mechanisms quickly becomes intractable. Consequently, empirical evaluation is used to further analyse the properties of the mechanisms. This analysis is first performed in an environment in which agents are able to manipulate their reports. However, further analysis is provided which shows the behaviour of the agents when they are able to make themselves harder to predict. Such a scenario has thus far not been discussed within mechanism design literature. Empirical analysis shows the sum of others' plus max mechanism to provide greater incentives for agents to make precise predictions. Furthermore, as a result of this, the aggregator increases its utility through implementing the sum of others' plus max mechanism over the uniform mechanism and over implementing no mechanism. Moreover, in settings which allow agents to manipulate the volatility of their loads, it is shown that the uniform mechanism causes the aggregator to lose utility in comparison to using no mechanism, whereas in comparison to no mechanism, the sum of others' plus max mechanism causes an increase in utility to the aggregator.
Rose, Harry
48024eb8-587d-423e-962f-c581a069921c
Rose, Harry
48024eb8-587d-423e-962f-c581a069921c
Rogers, Alexander
f9130bc6-da32-474e-9fab-6c6cb8077fdc

Rose, Harry (2013) Mechanism design for information aggregation within the smart grid. University of Southampton, Faculty of Physical and Applied Science, Doctoral Thesis, 157pp.

Record type: Thesis (Doctoral)

Abstract

The introduction of a smart electricity grid enables a greater amount of information exchange between consumers and their suppliers. This can be exploited by novel aggregation services to save money by more optimally purchasing electricity for those consumers. Now, if the aggregation service pays consumers for said information, then both parties could benefit. However, any such payment mechanism must be carefully designed to encourage the customers (say, home-owners) to exert effort in gathering information and then to truthfully report it to the aggregator. This work develops a model of the information aggregation problem where each home has an autonomous home agent, which acts on its behalf to gather information and report it to the aggregation agent. The aggregator has its own historical consumption information for each house under its service, so it can make an imprecise estimate of the future aggregate consumption of the houses for which it is responsible. However, it uses the information sent by the home agents in order to make a more precise estimate and, in return, gives each home agent a reward whose amount is determined by the payment mechanism in use by the aggregator. There are three desirable properties of a mechanism that this work considers: budget balance (the aggregator does not reward the agents more than it saves), incentive compatibility (agents are encouraged to report truthfully), and finally individual rationality (the payments to the home agents must outweigh their incurred costs). In this thesis, mechanism design is used to develop and analyse two mechanisms. The first, named the uniform mechanism, divides the savings made by the aggregator equally among the houses. This is both Nash incentive compatible, strongly budget balanced and individually rational. However, the agents' rewards are not fair insofar as each agent is rewarded equally irrespective of that agent's actual contribution to the savings. This results in a smaller incentive for agents to produce precise reports. Moreover, it encourages undesirable behaviour from agents who are able to make the loads placed upon the grid more volatile such that they are harder to predict. To resolve these issues, a novel scoring rule-based mechanism named sum of others' plus max is developed, which uses the spherical scoring rule to more fairly distribute rewards to agents based on the accuracy and precision of their individual reports. This mechanism is weakly budget balanced, dominant strategy incentive compatible and individually rational. Moreover, it encourages agents to make their loads less volatile, such that they are more predictable. This has obvious advantages to the electricity grid. For example, the amount of spinning reserve generation can be reduced, reducing the carbon output of the grid and the cost per unit of electricity. This work makes use of both theoretical and empirical analysis in order to evaluate the aforementioned mechanisms. Theoretical analysis is used in order to prove budget balance, individual rationality and incentive compatibility. However, theoretical evaluation of the equilibrium strategies within each of the mechanisms quickly becomes intractable. Consequently, empirical evaluation is used to further analyse the properties of the mechanisms. This analysis is first performed in an environment in which agents are able to manipulate their reports. However, further analysis is provided which shows the behaviour of the agents when they are able to make themselves harder to predict. Such a scenario has thus far not been discussed within mechanism design literature. Empirical analysis shows the sum of others' plus max mechanism to provide greater incentives for agents to make precise predictions. Furthermore, as a result of this, the aggregator increases its utility through implementing the sum of others' plus max mechanism over the uniform mechanism and over implementing no mechanism. Moreover, in settings which allow agents to manipulate the volatility of their loads, it is shown that the uniform mechanism causes the aggregator to lose utility in comparison to using no mechanism, whereas in comparison to no mechanism, the sum of others' plus max mechanism causes an increase in utility to the aggregator.

PDF
RoseThesis.pdf - Other
Download (2MB)

More information

Published date: February 2013
Organisations: University of Southampton, Agents, Interactions & Complexity

Identifiers

Local EPrints ID: 349119
URI: http://eprints.soton.ac.uk/id/eprint/349119
PURE UUID: 632f6f63-c8dd-470b-90d6-cbedf4b8f326

Catalogue record

Date deposited: 07 Mar 2013 14:33
Last modified: 18 Jul 2017 04:44

Export record

Contributors

Author: Harry Rose
Thesis advisor: Alexander Rogers

University divisions

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×