2015
Alam, Muddasser; Gerding, Enrico H.; Rogers, Alex; Ramchurn, Sarvapali D.
A scalable, decentralised multi-issue negotiation protocol for energy exchange Proceedings Article
In: International Joint Conference on Artificial Intelligence (IJCAI), 2015.
@inproceedings{eps376618,
title = {A scalable, decentralised multi-issue negotiation protocol for energy exchange},
author = {Muddasser Alam and Enrico H. Gerding and Alex Rogers and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/376618/},
year = {2015},
date = {2015-01-01},
booktitle = {International Joint Conference on Artificial Intelligence (IJCAI)},
abstract = {We present a novel negotiation protocol to facilitate energy exchange between off-grid homes that are equipped with renewable energy generation and electricity storage. Our protocol imposes restrictions over negotiation such that it reduces the complex interdependent multi-issue negotiation to one where agents have a strategy profile in subgame perfect Nash equilibrium. We show that our protocol is concurrent, scalable and; under certain conditions; leads to Pareto-optimal outcomes.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a novel negotiation protocol to facilitate energy exchange between off-grid homes that are equipped with renewable energy generation and electricity storage. Our protocol imposes restrictions over negotiation such that it reduces the complex interdependent multi-issue negotiation to one where agents have a strategy profile in subgame perfect Nash equilibrium. We show that our protocol is concurrent, scalable and; under certain conditions; leads to Pareto-optimal outcomes.
Wu, Feng; Ramchurn, Sarvapali; Jiang, Wenchao; Fischer, Joel; Rodden, Tom; Jennings, Nicholas R.
Agile Planning for Real-World Disaster Response Proceedings Article
In: International Joint Conference on Artificial Intelligence, 2015.
@inproceedings{eps377186,
title = {Agile Planning for Real-World Disaster Response},
author = {Feng Wu and Sarvapali Ramchurn and Wenchao Jiang and Joel Fischer and Tom Rodden and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/377186/},
year = {2015},
date = {2015-01-01},
booktitle = {International Joint Conference on Artificial Intelligence},
abstract = {We consider a setting where an agent-based planner instructs teams of human emergency responders to perform tasks in the real world. Due to uncertainty in the environment and the inability of the planner to consider all human preferences and all attributes of the real-world, humans may reject plans computed by the agent. A na??ve solution that replans given a rejection is inefficient and does not guarantee the new plan will be acceptable. Hence, we propose a new model re-planning problem using a Multi-agent Markov Decision Process that integrates potential rejections as part of the planning process and propose a novel algorithm to efficiently solve this new model. We empirically evaluate our algorithm and show that it outperforms current benchmarks. Our algorithm is also shown to perform better in pilot studies with real humans.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We consider a setting where an agent-based planner instructs teams of human emergency responders to perform tasks in the real world. Due to uncertainty in the environment and the inability of the planner to consider all human preferences and all attributes of the real-world, humans may reject plans computed by the agent. A na??ve solution that replans given a rejection is inefficient and does not guarantee the new plan will be acceptable. Hence, we propose a new model re-planning problem using a Multi-agent Markov Decision Process that integrates potential rejections as part of the planning process and propose a novel algorithm to efficiently solve this new model. We empirically evaluate our algorithm and show that it outperforms current benchmarks. Our algorithm is also shown to perform better in pilot studies with real humans.
Ramchurn, Sarvapali; Fischer, Joel; Ikuno, Yuki; Wu, Feng; Flann, Jack; Waldock, Antony
A Study of Human-Agent Collaboration for Multi-UAV Task Allocation in Dynamic Environments Proceedings Article
In: International Joint Conference on Artificial Intelligence, 2015.
@inproceedings{eps377185,
title = {A Study of Human-Agent Collaboration for Multi-UAV Task Allocation in Dynamic Environments},
author = {Sarvapali Ramchurn and Joel Fischer and Yuki Ikuno and Feng Wu and Jack Flann and Antony Waldock},
url = {http://eprints.soton.ac.uk/377185/},
year = {2015},
date = {2015-01-01},
booktitle = {International Joint Conference on Artificial Intelligence},
abstract = {We consider a setting where a team of humans oversee the coordination of multiple Unmanned Aerial Vehicles (UAVs) to perform a number of search tasks in dynamic environments that may cause the UAVs to drop out. Hence, we develop a set of multi-UAV supervisory control interfaces and a multi-agent coordination algorithm to support human decision making in this setting. To elucidate the resulting interactional issues, we compare manual and mixed-initiative task allocation in both static and dynamic environments in lab studies with 40 participants and observe that our mixed-initiative system results in lower workloads and better performance in re-planning tasks than one which only involves manual task allocation. Our analysis points to new insights into the way humans appropriate flexible autonomy.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We consider a setting where a team of humans oversee the coordination of multiple Unmanned Aerial Vehicles (UAVs) to perform a number of search tasks in dynamic environments that may cause the UAVs to drop out. Hence, we develop a set of multi-UAV supervisory control interfaces and a multi-agent coordination algorithm to support human decision making in this setting. To elucidate the resulting interactional issues, we compare manual and mixed-initiative task allocation in both static and dynamic environments in lab studies with 40 participants and observe that our mixed-initiative system results in lower workloads and better performance in re-planning tasks than one which only involves manual task allocation. Our analysis points to new insights into the way humans appropriate flexible autonomy.
Holyhead, James C.; Ramchurn, Sarvapali D.; Rogers, Alex
Consumer Targeting in Residential Demand Response Programmes Proceedings Article
In: Proceedings of the 2015 ACM Sixth International Conference on Future Energy Systems, pp. 7–16, ACM, Bangalore, India, 2015, ISBN: 978-1-4503-3609-3.
@inproceedings{Holyhead:2015:CTR:2768510.2768531,
title = {Consumer Targeting in Residential Demand Response Programmes},
author = {Holyhead, James C. and Ramchurn, Sarvapali D. and Rogers, Alex},
url = {http://doi.acm.org/10.1145/2768510.2768531},
isbn = {978-1-4503-3609-3},
year = {2015},
date = {2015-01-01},
booktitle = {Proceedings of the 2015 ACM Sixth International Conference on Future Energy Systems},
pages = {7--16},
publisher = {ACM},
address = {Bangalore, India},
series = {e-Energy '15},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Chen, Shaofei; Wu, Feng; Shen, Lincheng; Chen, Jing; Ramchurn, Sarvapali D
Multi-Agent Patrolling under Uncertainty and Threats Journal Article
In: PLoS ONE, vol. 10, no. 6, pp. e0130154, 2015, ISBN: 1932-6203.
@article{chen:etal:2016,
title = {Multi-Agent Patrolling under Uncertainty and Threats},
author = {Chen, Shaofei and Wu, Feng and Shen, Lincheng and Chen, Jing and Ramchurn, Sarvapali D},
editor = {Deng, Yong},
url = {http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4472811/},
doi = {10.1371/journal.pone.0130154},
isbn = {1932-6203},
year = {2015},
date = {2015-01-01},
journal = {PLoS ONE},
volume = {10},
number = {6},
pages = {e0130154},
publisher = {Public Library of Science},
abstract = {We investigate a multi-agent patrolling problem where information is distributed alongside threats in environments with uncertainties. Specifically, the information and threat at each location are independently modelled as multi-state Markov chains, whose states are not observed until the location is visited by an agent. While agents will obtain information at a location, they may also suffer damage from the threat at that location. Therefore, the goal of the agents is to gather as much information as possible while mitigating the damage incurred. To address this challenge, we formulate the single-agent patrolling problem as a Partially Observable Markov Decision Process (POMDP) and propose a computationally efficient algorithm to solve this model. Building upon this, to compute patrols for multiple agents, the single-agent algorithm is extended for each agent with the aim of maximising its marginal contribution to the team. We empirically evaluate our algorithm on problems of multi-agent patrolling and show that it outperforms a baseline algorithm up to 44% for 10 agents and by 21% for 15 agents in large domains.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We investigate a multi-agent patrolling problem where information is distributed alongside threats in environments with uncertainties. Specifically, the information and threat at each location are independently modelled as multi-state Markov chains, whose states are not observed until the location is visited by an agent. While agents will obtain information at a location, they may also suffer damage from the threat at that location. Therefore, the goal of the agents is to gather as much information as possible while mitigating the damage incurred. To address this challenge, we formulate the single-agent patrolling problem as a Partially Observable Markov Decision Process (POMDP) and propose a computationally efficient algorithm to solve this model. Building upon this, to compute patrols for multiple agents, the single-agent algorithm is extended for each agent with the aim of maximising its marginal contribution to the team. We empirically evaluate our algorithm on problems of multi-agent patrolling and show that it outperforms a baseline algorithm up to 44% for 10 agents and by 21% for 15 agents in large domains.
Chen, S.; Wu, F.; Shen, L.; Chen, J.; Ramchurn, S. D.
Decentralized Patrolling Under Constraints in Dynamic Environments Journal Article
In: Cybernetics, IEEE Transactions on, vol. PP, no. 99, pp. 1-13, 2015, ISSN: 2168-2267.
@article{7362160,
title = {Decentralized Patrolling Under Constraints in Dynamic Environments},
author = {Chen, S. and Wu, F. and Shen, L. and Chen, J. and Ramchurn, S.D.},
doi = {10.1109/TCYB.2015.2505737},
issn = {2168-2267},
year = {2015},
date = {2015-01-01},
journal = {Cybernetics, IEEE Transactions on},
volume = {PP},
number = {99},
pages = {1-13},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Kalyanaraman, Shivkumar; Seetharam, Deva P; Shorey, Rajeev; Ramchurn, Sarvapali D; Srivastava, Mani (Ed.)
ACM, 2015, ISBN: 978-1-4503-3609-3.
@proceedings{DBLP:conf/eenergy/2015,
title = {Proceedings of the 2015 ACM Sixth International Conference on Future
Energy Systems, e-Energy 2015, Bangalore, India, July 14-17, 2015},
editor = {Shivkumar Kalyanaraman and Deva P Seetharam and Rajeev Shorey and Sarvapali D Ramchurn and Mani Srivastava},
url = {http://dl.acm.org/citation.cfm?id=2768510},
isbn = {978-1-4503-3609-3},
year = {2015},
date = {2015-01-01},
publisher = {ACM},
keywords = {},
pubstate = {published},
tppubtype = {proceedings}
}
2014
F. Recchia M. Bicego, A. Farinelli
Behavioural biometrics using electricity load profiles Journal Article
In: Proceedings of the International Conference on Pattern Recognition, 2014.
@article{bicego:etal:2014,
title = {Behavioural biometrics using electricity load profiles},
author = {M. Bicego, F. Recchia, A. Farinelli, S. D. Ramchurn, E. Grosso},
url = {http://www.sramchurn.com/wp-content/uploads/2014/10/CR_v1.pdf},
year = {2014},
date = {2014-08-24},
journal = {Proceedings of the International Conference on Pattern Recognition},
abstract = {Modelling behavioural biometric patterns is a key
issue for modern user centric applications, aimed at better monitoring
users’ activities, understanding their habits and detecting
their identity. Following this trend, this paper investigates whether
the electrical energy consumption of a user can be a distinctive
behavioural biometric trait. In particular we analyse daily and
weekly load profiles showing that they are closely related to
the identity of the users. Hence, we believe that this level of
analysis can open interesting application scenarios in the field of
energy management and it provides a good working framework
for the continuous development of smart environments with
demonstrable benefits on real-world implementations.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Modelling behavioural biometric patterns is a key
issue for modern user centric applications, aimed at better monitoring
users’ activities, understanding their habits and detecting
their identity. Following this trend, this paper investigates whether
the electrical energy consumption of a user can be a distinctive
behavioural biometric trait. In particular we analyse daily and
weekly load profiles showing that they are closely related to
the identity of the users. Hence, we believe that this level of
analysis can open interesting application scenarios in the field of
energy management and it provides a good working framework
for the continuous development of smart environments with
demonstrable benefits on real-world implementations.
issue for modern user centric applications, aimed at better monitoring
users’ activities, understanding their habits and detecting
their identity. Following this trend, this paper investigates whether
the electrical energy consumption of a user can be a distinctive
behavioural biometric trait. In particular we analyse daily and
weekly load profiles showing that they are closely related to
the identity of the users. Hence, we believe that this level of
analysis can open interesting application scenarios in the field of
energy management and it provides a good working framework
for the continuous development of smart environments with
demonstrable benefits on real-world implementations.
Alan, Alper; Costanza, Enrico; Fischer, J.; Ramchurn, Sarvapali; Rodden, T.; Jennings, N. R.
A field study of human-agent interaction for electricity tariff switching Proceedings Article
In: Proceedings of the 13th International Conference on Autonomous Agents and Multi-Agent Systems, 2014.
@inproceedings{eps360820,
title = {A field study of human-agent interaction for electricity tariff switching},
author = {Alper Alan and Enrico Costanza and J. Fischer and Sarvapali Ramchurn and T. Rodden and N.R. Jennings},
url = {http://eprints.soton.ac.uk/360820/},
year = {2014},
date = {2014-01-01},
booktitle = {Proceedings of the 13th International Conference on Autonomous Agents and Multi-Agent Systems},
abstract = {Recently, many algorithms have been developed for autonomous agents to manage home energy use on behalf of their human owners. By so doing, it is expected that agents will be more efficient at, for example, choosing the best energy tariff to switch to when dynamically priced tariffs come about. However, to date, there has been no validation of such technologies in any field trial. In particular, it has not been shown whether users prefer fully autonomous agents as opposed to controlling their preferences manually. Hence, in this paper we describe a novel platform, called Tariff Agent, to study notions of flexible autonomy in the context of tariff switching. Tariff Agent uses real-world datasets and real-time electricity monitoring to instantiate a scenario where human participants may have to make, or delegate to their agent (in different ways), tariff switching decisions given uncertainties about their own consumption and tariff prices. We carried out a field trial with 10 participants and, from both quantitative and qualitative results, formulate novel design guidelines for systems that implement flexible autonom.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Recently, many algorithms have been developed for autonomous agents to manage home energy use on behalf of their human owners. By so doing, it is expected that agents will be more efficient at, for example, choosing the best energy tariff to switch to when dynamically priced tariffs come about. However, to date, there has been no validation of such technologies in any field trial. In particular, it has not been shown whether users prefer fully autonomous agents as opposed to controlling their preferences manually. Hence, in this paper we describe a novel platform, called Tariff Agent, to study notions of flexible autonomy in the context of tariff switching. Tariff Agent uses real-world datasets and real-time electricity monitoring to instantiate a scenario where human participants may have to make, or delegate to their agent (in different ways), tariff switching decisions given uncertainties about their own consumption and tariff prices. We carried out a field trial with 10 participants and, from both quantitative and qualitative results, formulate novel design guidelines for systems that implement flexible autonom.
Costanza, Enrico; Fischer, Joel E; Colley, James A; Rodden, Tom; Ramchurn, Sarvapali; Jennings, Nicholas R.
Doing the laundry with agents: a field trial of a future smart energy system in the home Proceedings Article
In: ACM CHI Conference on Human Factors in Computing Systems 2014, pp. 813–822, ACM 2014.
@inproceedings{eps361173,
title = {Doing the laundry with agents: a field trial of a future smart energy system in the home},
author = {Enrico Costanza and Joel E Fischer and James A Colley and Tom Rodden and Sarvapali Ramchurn and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/361173/},
year = {2014},
date = {2014-01-01},
booktitle = {ACM CHI Conference on Human Factors in Computing Systems 2014},
pages = {813--822},
organization = {ACM},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Tran-Thanh, Long; Huynh, Trung Dong; Rosenfield, A; Ramchurn, Sarvapali; Jennings, Nicholas R.
BudgetFix: budget limited crowdsourcing for interdependent task allocation with quality guarantees Proceedings Article
In: 13th International Conference on Autonomous Agents and Multi-Agent Systems, International Foundation for Autonomous Agents and Multiagent Systems, 2014.
@inproceedings{eps362321,
title = {BudgetFix: budget limited crowdsourcing for interdependent task allocation with quality guarantees},
author = {Long Tran-Thanh and Trung Dong Huynh and A Rosenfield and Sarvapali Ramchurn and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/362321/},
year = {2014},
date = {2014-01-01},
booktitle = {13th International Conference on Autonomous Agents and Multi-Agent Systems},
publisher = {International Foundation for Autonomous Agents and Multiagent Systems},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Vinyals, Meritxell; Macarthur, Kathryn; Farinelli, Alessandro; Ramchurn, Sarvapali; Jennings, Nicholas R.
A message-passing approach to decentralised parallel machine scheduling Journal Article
In: The Computer Journal, 2014.
@article{eps360818,
title = {A message-passing approach to decentralised parallel machine scheduling},
author = {Meritxell Vinyals and Kathryn Macarthur and Alessandro Farinelli and Sarvapali Ramchurn and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/360818/},
year = {2014},
date = {2014-01-01},
journal = {The Computer Journal},
publisher = {Oxford University Press},
abstract = {This paper tackles the problem of parallelizing heterogeneous computational tasks across a number of computational nodes (aka agents) where each agent may not be able to perform all the tasks and may have different computational speeds. An equivalent problem can be found in operations research, and it is known as scheduling tasks on unrelated parallel machines (also known as R?Cmax). Given this equivalence observation, we present the spanning tree decentralized task distribution algorithm (ST-DTDA), the first decentralized solution to R?Cmax. ST-DTDA achieves decomposition by means of the min?max algorithm, a member of the generalized distributive law family, that performs inference by message-passing along the edges of a graphical model (known as a junction tree). Specifically, ST-DTDA uses min?max to optimally solve an approximation of the original R?Cmax problem that results from eliminating possible agent-task allocations until it is mapped into an acyclic structure. To eliminate those allocations that are least likely to have an impact on the solution quality, ST-DTDA uses a heuristic approach. Moreover, ST-DTDA provides a per-instance approximation ratio that guarantees that the makespan of its solution (optimal in the approximated R?Cmax problem) is not more than a factor ensuremathrho times the makespan of the optimal of the original problem. In our empirical evaluation of ST-DTDA, we show that ST-DTDA, with a min-regret heuristic, converges to solutions that are between 78 and 95% optimal whilst providing approximation ratios lower than 3.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper tackles the problem of parallelizing heterogeneous computational tasks across a number of computational nodes (aka agents) where each agent may not be able to perform all the tasks and may have different computational speeds. An equivalent problem can be found in operations research, and it is known as scheduling tasks on unrelated parallel machines (also known as R?Cmax). Given this equivalence observation, we present the spanning tree decentralized task distribution algorithm (ST-DTDA), the first decentralized solution to R?Cmax. ST-DTDA achieves decomposition by means of the min?max algorithm, a member of the generalized distributive law family, that performs inference by message-passing along the edges of a graphical model (known as a junction tree). Specifically, ST-DTDA uses min?max to optimally solve an approximation of the original R?Cmax problem that results from eliminating possible agent-task allocations until it is mapped into an acyclic structure. To eliminate those allocations that are least likely to have an impact on the solution quality, ST-DTDA uses a heuristic approach. Moreover, ST-DTDA provides a per-instance approximation ratio that guarantees that the makespan of its solution (optimal in the approximated R?Cmax problem) is not more than a factor ensuremathrho times the makespan of the optimal of the original problem. In our empirical evaluation of ST-DTDA, we show that ST-DTDA, with a min-regret heuristic, converges to solutions that are between 78 and 95% optimal whilst providing approximation ratios lower than 3.
Fischer, J. E.; Jiang, W; Kerne, A; Greenhalgh, C; Ramchurn, Sarvapali D; Reece, Steven; Pantidi, N; Rodden, T
Supporting Team Coordination on the Ground: Requirements from a Mixed Reality Game. Proceedings Article
In: 11th Int. Conference on the Design of Cooperative Systems (COOP ?14), 2014.
@inproceedings{orchid192,
title = {Supporting Team Coordination on the Ground: Requirements from a Mixed Reality Game.},
author = {J.E. Fischer and W Jiang and A Kerne and C Greenhalgh and Sarvapali D Ramchurn and Steven Reece and N Pantidi and T Rodden},
year = {2014},
date = {2014-01-01},
booktitle = {11th Int. Conference on the Design of Cooperative Systems (COOP ?14)},
howpublished = {http://www.orchid.ac.uk/eprints/192/1/COOP2014-Fischer-author-version.pdf},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Jiang, W; Fischer, J. E.; Greenhalgh, C; Ramchurn, Sarvapali D; Wu, Feng; Jennings, Nicholas R; Rodden, T
Social Implications of Agent-based Planning Support for Human Teams. Proceedings Article
In: 2014 Int. Conference on Collaboration Technologies and Systems, 2014.
@inproceedings{orchid191,
title = {Social Implications of Agent-based Planning Support for Human Teams.},
author = {W Jiang and J.E. Fischer and C Greenhalgh and Sarvapali D Ramchurn and Feng Wu and Nicholas R Jennings and T Rodden},
year = {2014},
date = {2014-01-01},
booktitle = {2014 Int. Conference on Collaboration Technologies and Systems},
howpublished = {http://www.orchid.ac.uk/eprints/191/1/CTS2014-Jiang-author-version.pdf},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Pawlowski, Krzysztof; Kurach, Karol; Svensson, Kim; Ramchurn, Sarvapali D; Michalak, Tomasz; Rahwan, Talal
Coalition Structure Generation with the Graphics Processing Unit Proceedings Article
In: 13th Int. Conf. on Autonomous Agents and Multi-Agent Systems, 2014.
@inproceedings{orchid176,
title = {Coalition Structure Generation with the Graphics Processing Unit},
author = {Krzysztof Pawlowski and Karol Kurach and Kim Svensson and Sarvapali D Ramchurn and Tomasz Michalak and Talal Rahwan},
year = {2014},
date = {2014-01-01},
booktitle = {13th Int. Conf. on Autonomous Agents and Multi-Agent Systems},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Huynh, Trung Dong; Ebden, Mark; Ramchurn, Sarvapali; Roberts, Stephen; Moreau, Luc
Data quality assessment from provenance graphs Proceedings Article
In: Provenance Analytics 2014, 2014.
@inproceedings{eps365510,
title = {Data quality assessment from provenance graphs},
author = {Trung Dong Huynh and Mark Ebden and Sarvapali Ramchurn and Stephen Roberts and Luc Moreau},
url = {http://eprints.soton.ac.uk/365510/},
year = {2014},
date = {2014-01-01},
booktitle = {Provenance Analytics 2014},
abstract = {Provenance is a domain-independent means to represent what happened in an application, which can help verify data and infer data quality. Provenance patterns can manifest real-world phenomena such as a significant interest in a piece of content, providing an indication of its quality, or even issues such as undesirable interactions within a group of contributors. This paper presents an application-independent methodology for analyzing data based on the network metrics of provenance graphs to learn about such patterns and to relate them to data quality in an automated manner. Validating this method on the provenance records of CollabMap, an online crowdsourcing mapping application, we demonstrated an accuracy level of over 95% for the trust classification of data generated by the crowd therein.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Provenance is a domain-independent means to represent what happened in an application, which can help verify data and infer data quality. Provenance patterns can manifest real-world phenomena such as a significant interest in a piece of content, providing an indication of its quality, or even issues such as undesirable interactions within a group of contributors. This paper presents an application-independent methodology for analyzing data based on the network metrics of provenance graphs to learn about such patterns and to relate them to data quality in an automated manner. Validating this method on the provenance records of CollabMap, an online crowdsourcing mapping application, we demonstrated an accuracy level of over 95% for the trust classification of data generated by the crowd therein.
Jennings, Nicholas R.; Moreau, Luc; Nicholson, D; Ramchurn, Sarvapali D.; Roberts, S; Rodden, T; Rogers, Alex
On human-agent collectives Journal Article
In: Communications of the ACM, vol. 57, no. 12, pp. 33-42, 2014.
@article{eps364593,
title = {On human-agent collectives},
author = {Nicholas R. Jennings and Luc Moreau and D Nicholson and Sarvapali D. Ramchurn and S Roberts and T Rodden and Alex Rogers},
url = {http://eprints.soton.ac.uk/364593/},
year = {2014},
date = {2014-01-01},
journal = {Communications of the ACM},
volume = {57},
number = {12},
pages = {33-42},
abstract = {We live in a world where a host of computer systems, distributed throughout our physical and information environments, are increasingly implicated in our everyday actions. Computer technologies impact all aspects of our lives and our relationship with the digital has fundamentally altered as computers have moved out of the workplace and away from the desktop. Networked computers, tablets, phones and personal devices are now commonplace, as are an increasingly diverse set of digital devices built into the world around us. Data and information is generated at unprecedented speeds and volumes from an increasingly diverse range of sources. It is then combined in unforeseen ways, limited only by human imagination. People?s activities and collaborations are becoming ever more dependent upon and intertwined with this ubiquitous information substrate. As these trends continue apace, it is becoming apparent that many endeavours involve the symbiotic interleaving of humans and computers. Moreover, the emergence of these close-knit partnerships is inducing profound change. Rather than issuing instructions to passive machines that wait until they are asked before doing anything, we will work in tandem with highly inter-connected computational components that act autonomously and intelligently (aka agents). As a consequence, greater attention needs to be given to the balance of control between people and machines. In many situations, humans will be in charge and agents will predominantly act in a supporting role. In other cases, however, the agents will be in control and humans will play the supporting role. We term this emerging class of systems human-agent collectives (HACs) to reflect the close partnership and the flexible social interactions between the humans and the computers. As well as exhibiting increased autonomy, such systems will be inherently open and social. This means the participants will need to continually and flexibly establish and manage a range of social relationships. Thus, depending on the task at hand, different constellations of people, resources, and information will need to come together, operate in a coordinated fashion, and then disband. The openness and presence of many distinct stakeholders means participation will be motivated by a broad range of incentives rather than diktat. This article outlines the key research challenges involved in developing a comprehensive understanding of HACs. To illuminate this agenda, a nascent application in the domain of disaster response is presented.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We live in a world where a host of computer systems, distributed throughout our physical and information environments, are increasingly implicated in our everyday actions. Computer technologies impact all aspects of our lives and our relationship with the digital has fundamentally altered as computers have moved out of the workplace and away from the desktop. Networked computers, tablets, phones and personal devices are now commonplace, as are an increasingly diverse set of digital devices built into the world around us. Data and information is generated at unprecedented speeds and volumes from an increasingly diverse range of sources. It is then combined in unforeseen ways, limited only by human imagination. People?s activities and collaborations are becoming ever more dependent upon and intertwined with this ubiquitous information substrate. As these trends continue apace, it is becoming apparent that many endeavours involve the symbiotic interleaving of humans and computers. Moreover, the emergence of these close-knit partnerships is inducing profound change. Rather than issuing instructions to passive machines that wait until they are asked before doing anything, we will work in tandem with highly inter-connected computational components that act autonomously and intelligently (aka agents). As a consequence, greater attention needs to be given to the balance of control between people and machines. In many situations, humans will be in charge and agents will predominantly act in a supporting role. In other cases, however, the agents will be in control and humans will play the supporting role. We term this emerging class of systems human-agent collectives (HACs) to reflect the close partnership and the flexible social interactions between the humans and the computers. As well as exhibiting increased autonomy, such systems will be inherently open and social. This means the participants will need to continually and flexibly establish and manage a range of social relationships. Thus, depending on the task at hand, different constellations of people, resources, and information will need to come together, operate in a coordinated fashion, and then disband. The openness and presence of many distinct stakeholders means participation will be motivated by a broad range of incentives rather than diktat. This article outlines the key research challenges involved in developing a comprehensive understanding of HACs. To illuminate this agenda, a nascent application in the domain of disaster response is presented.
Bistaffa, Filippo; Farinelli, Alessandro; Cerquides, Jesus; Rodriguez-Aguilar, Juan Antonio; Ramchurn, Sarvapali D
Anytime Coalition Structure Generation on Synergy Graphs Proceedings Article
In: 13th Int. Conf. on Autonomous Agents and Multi-Agent Systems, pp. 13-20, 2014.
@inproceedings{orchid175,
title = {Anytime Coalition Structure Generation on Synergy Graphs},
author = {Filippo Bistaffa and Alessandro Farinelli and Jesus Cerquides and Juan Antonio Rodriguez-Aguilar and Sarvapali D Ramchurn},
url = {http://aamas2014.lip6.fr/proceedings/aamas/p13.pdf},
year = {2014},
date = {2014-01-01},
booktitle = {13th Int. Conf. on Autonomous Agents and Multi-Agent Systems},
pages = {13-20},
abstract = {We consider the coalition structure generation (CSG) problem on
synergy graphs, which arises in many practical applications where
communication constraints, social or trust relationships must be
taken into account when forming coalitions. We propose a novel
representation of this problem based on the concept of edge contraction,
and an innovative branch and bound approach (CFSS),
which is particularly efficient when applied to a general class of
characteristic functions. This new model provides a non-redundant
partition of the search space, hence allowing an effective parallelisation.
We evaluate CFSS on two benchmark functions, the edge
sum with coordination cost and the collective energy purchasing
functions, comparing its performance with the best algorithm for
CSG on synergy graphs: DyCE. The latter approach is centralised
and cannot be efficiently parallelised due to the exponential memory
requirements in the number of agents, which limits its scalability
(while CFSS memory requirements are only polynomial).
Our results show that, when the graphs are very sparse, CFSS is
4 orders of magnitude faster than DyCE. Moreover, CFSS is the
first approach to provide anytime approximate solutions with quality
guarantees for very large systems (i.e., with more than 2700
agents).},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We consider the coalition structure generation (CSG) problem on
synergy graphs, which arises in many practical applications where
communication constraints, social or trust relationships must be
taken into account when forming coalitions. We propose a novel
representation of this problem based on the concept of edge contraction,
and an innovative branch and bound approach (CFSS),
which is particularly efficient when applied to a general class of
characteristic functions. This new model provides a non-redundant
partition of the search space, hence allowing an effective parallelisation.
We evaluate CFSS on two benchmark functions, the edge
sum with coordination cost and the collective energy purchasing
functions, comparing its performance with the best algorithm for
CSG on synergy graphs: DyCE. The latter approach is centralised
and cannot be efficiently parallelised due to the exponential memory
requirements in the number of agents, which limits its scalability
(while CFSS memory requirements are only polynomial).
Our results show that, when the graphs are very sparse, CFSS is
4 orders of magnitude faster than DyCE. Moreover, CFSS is the
first approach to provide anytime approximate solutions with quality
guarantees for very large systems (i.e., with more than 2700
agents).
synergy graphs, which arises in many practical applications where
communication constraints, social or trust relationships must be
taken into account when forming coalitions. We propose a novel
representation of this problem based on the concept of edge contraction,
and an innovative branch and bound approach (CFSS),
which is particularly efficient when applied to a general class of
characteristic functions. This new model provides a non-redundant
partition of the search space, hence allowing an effective parallelisation.
We evaluate CFSS on two benchmark functions, the edge
sum with coordination cost and the collective energy purchasing
functions, comparing its performance with the best algorithm for
CSG on synergy graphs: DyCE. The latter approach is centralised
and cannot be efficiently parallelised due to the exponential memory
requirements in the number of agents, which limits its scalability
(while CFSS memory requirements are only polynomial).
Our results show that, when the graphs are very sparse, CFSS is
4 orders of magnitude faster than DyCE. Moreover, CFSS is the
first approach to provide anytime approximate solutions with quality
guarantees for very large systems (i.e., with more than 2700
agents).
2013
Alam, Muddasser; Alan, Alper; Rogers, Alex; Ramchurn, Sarvapali D.
Towards a smart home framework Proceedings Article
In: 5th ACM Workshop On Embedded Systems For Energy-Efficient Buildings (BuildSys2013), 2013.
@inproceedings{eps357187,
title = {Towards a smart home framework},
author = {Muddasser Alam and Alper Alan and Alex Rogers and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/357187/},
year = {2013},
date = {2013-01-01},
booktitle = {5th ACM Workshop On Embedded Systems For Energy-Efficient Buildings (BuildSys2013)},
abstract = {We present our Smart Home Framework (SHF) which simplifies the modelling, prototyping and simulation of smart infrastructure (i.e., smart home and smart communities). It provides the buildings blocks (e.g., home appliances) that can be extended and assembled together to build a smart infrastructure model to which appropriate AI techniques can be applied. This approach enables rapid modelling where new research initiatives can build on existing work.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present our Smart Home Framework (SHF) which simplifies the modelling, prototyping and simulation of smart infrastructure (i.e., smart home and smart communities). It provides the buildings blocks (e.g., home appliances) that can be extended and assembled together to build a smart infrastructure model to which appropriate AI techniques can be applied. This approach enables rapid modelling where new research initiatives can build on existing work.
Alam, Muddasser; Ramchurn, Sarvapali; Rogers, Alex
Cooperative energy exchange for the efficient use of energy and resources in remote communities. [Winner, Best Student Paper Award at AAMAS2013] Proceedings Article
In: Twelfth International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2013), 2013.
@inproceedings{eps346637,
title = {Cooperative energy exchange for the efficient use of energy and resources in remote communities. [Winner, Best Student Paper Award at AAMAS2013]},
author = {Muddasser Alam and Sarvapali Ramchurn and Alex Rogers},
url = {http://eprints.soton.ac.uk/346637/},
year = {2013},
date = {2013-01-01},
booktitle = {Twelfth International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2013)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Alam, Muddasser; Rogers, Alex; Ramchurn, Sarvapali
Interdependent multi-issue negotiation for energy exchange in remote communities Proceedings Article
In: Twenty-Seventh AAAI Conference on Artificial Intelligence (AAAI-13), 2013.
@inproceedings{eps350941,
title = {Interdependent multi-issue negotiation for energy exchange in remote communities},
author = {Muddasser Alam and Alex Rogers and Sarvapali Ramchurn},
url = {http://eprints.soton.ac.uk/350941/},
year = {2013},
date = {2013-01-01},
booktitle = {Twenty-Seventh AAAI Conference on Artificial Intelligence (AAAI-13)},
abstract = {We present a novel negotiation protocol to facilitate energy exchange between off-grid homes that are equipped with renewable energy generation and electricity storage. Our protocol imposes restrictions over negotiation such that it reduces the complex interdependent multi-issue negotiation to one where agents have a strategy profile in subgame perfect Nash equilibrium. We show that our negotiation protocol is tractable, concurrent, scalable and leads to Pareto-optimal outcomes in a decentralised manner. We empirically evaluate our protocol and show that, in this instance, a society of agents can (i) improve the overall utilities by 14% and (ii) reduce their overall use of the batteries by 37%},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a novel negotiation protocol to facilitate energy exchange between off-grid homes that are equipped with renewable energy generation and electricity storage. Our protocol imposes restrictions over negotiation such that it reduces the complex interdependent multi-issue negotiation to one where agents have a strategy profile in subgame perfect Nash equilibrium. We show that our negotiation protocol is tractable, concurrent, scalable and leads to Pareto-optimal outcomes in a decentralised manner. We empirically evaluate our protocol and show that, in this instance, a society of agents can (i) improve the overall utilities by 14% and (ii) reduce their overall use of the batteries by 37%
Alam, Muddasser; Rogers, Alex; Ramchurn, Sarvapali D.
Interdependent multi-issue negotiation for energy exchange in remote communities Proceedings Article
In: International Workshop on AI Problems and Approaches for Intelligent Environments (AI4IE), 2013.
@inproceedings{eps357186,
title = {Interdependent multi-issue negotiation for energy exchange in remote communities},
author = {Muddasser Alam and Alex Rogers and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/357186/},
year = {2013},
date = {2013-01-01},
booktitle = {International Workshop on AI Problems and Approaches for Intelligent Environments (AI4IE)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Cerquides, Jesus; Farinelli, Alessandro; Meseguer, Pedro; Ramchurn, Sarvapali
A tutorial on optimisation for multi-agent systems Journal Article
In: The Computer Journal, pp. 1–26, 2013.
@article{eps361998,
title = {A tutorial on optimisation for multi-agent systems},
author = {Jesus Cerquides and Alessandro Farinelli and Pedro Meseguer and Sarvapali Ramchurn},
url = {http://eprints.soton.ac.uk/361998/},
year = {2013},
date = {2013-01-01},
journal = {The Computer Journal},
pages = {1--26},
abstract = {Research on optimization in multi-agent systems (MASs) has contributed with a wealth of techniques to solve many of the challenges arising in a wide range of multi-agent application domains. Multi-agent optimization focuses on casting MAS problems into optimization problems. The solving of those problems could possibly involve the active participation of the agents in a MAS. Research on multi-agent optimization has rapidly become a very technical, specialized field. Moreover, the contributions to the field in the literature are largely scattered. These two factors dramatically hinder access to a basic, general view of the foundations of the field. This tutorial is intended to ease such access by providing a gentle introduction to fundamental concepts and techniques on multi-agent optimization},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Research on optimization in multi-agent systems (MASs) has contributed with a wealth of techniques to solve many of the challenges arising in a wide range of multi-agent application domains. Multi-agent optimization focuses on casting MAS problems into optimization problems. The solving of those problems could possibly involve the active participation of the agents in a MAS. Research on multi-agent optimization has rapidly become a very technical, specialized field. Moreover, the contributions to the field in the literature are largely scattered. These two factors dramatically hinder access to a basic, general view of the foundations of the field. This tutorial is intended to ease such access by providing a gentle introduction to fundamental concepts and techniques on multi-agent optimization
Farinelli, Alessandro; Bicego, Manuele; Ramchurn, Sarvapali; Zuchelli, Marco
C-Link: a hierarchical clustering approach to large-scale near-optimal coalition formation Proceedings Article
In: 23rd International Joint Conference on Artificial Intelligence, AAAI Press / International Joint Conferences on Artificial Intelligence, 2013.
@inproceedings{eps351521,
title = {C-Link: a hierarchical clustering approach to large-scale near-optimal coalition formation},
author = {Alessandro Farinelli and Manuele Bicego and Sarvapali Ramchurn and Marco Zuchelli},
url = {http://eprints.soton.ac.uk/351521/},
year = {2013},
date = {2013-01-01},
booktitle = {23rd International Joint Conference on Artificial Intelligence},
publisher = {AAAI Press / International Joint Conferences on Artificial Intelligence},
abstract = {Coalition formation is a fundamental approach to multi-agent coordination. In this paper we address the specific problem of coalition structure generation, and focus on providing good-enough solutions using a novel heuristic approach that is based on data clustering methods. In particular, we propose a hierarchical agglomerative clustering approach (C-Link), which uses a similarity criterion between coalitions based on the gain that the system achieves if two coalitions merge. We empirically evaluate C-Link on a synthetic benchmark data-set as well as in collective energy purchasing settings. Our results show that the C-link approach performs very well against an optimal benchmark based on Mixed-Integer Programming, achieving solutions which are in the worst case about 80% of the optimal (in the synthetic data-set), and 98% of the optimal (in the energy data-set). Thus we show that C-Link can return solutions for problems involving thousands of agents within minutes.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Coalition formation is a fundamental approach to multi-agent coordination. In this paper we address the specific problem of coalition structure generation, and focus on providing good-enough solutions using a novel heuristic approach that is based on data clustering methods. In particular, we propose a hierarchical agglomerative clustering approach (C-Link), which uses a similarity criterion between coalitions based on the gain that the system achieves if two coalitions merge. We empirically evaluate C-Link on a synthetic benchmark data-set as well as in collective energy purchasing settings. Our results show that the C-link approach performs very well against an optimal benchmark based on Mixed-Integer Programming, achieving solutions which are in the worst case about 80% of the optimal (in the synthetic data-set), and 98% of the optimal (in the energy data-set). Thus we show that C-Link can return solutions for problems involving thousands of agents within minutes.
Fischer, Joel E.; Ramchurn, Sarvapali D.; Osborne, Michael A.; Parson, Oliver; Huynh, Trung Dong; Alam, Muddasser; Pantidi, Nadia; Moran, Stuart; Bachour, Khaled; Reece, Steven; Costanza, Enrico; Rodden, Tom; Jennings, Nicholas R.
Recommending Energy Tariffs and Load Shifting Based on Smart Household Usage Profiling Proceedings Article
In: International Conference on Intelligent User Interfaces, pp. 383–394, 2013.
@inproceedings{eps346991,
title = {Recommending Energy Tariffs and Load Shifting Based on Smart Household Usage Profiling},
author = {Joel E. Fischer and Sarvapali D. Ramchurn and Michael A. Osborne and Oliver Parson and Trung Dong Huynh and Muddasser Alam and Nadia Pantidi and Stuart Moran and Khaled Bachour and Steven Reece and Enrico Costanza and Tom Rodden and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/346991/},
year = {2013},
date = {2013-01-01},
booktitle = {International Conference on Intelligent User Interfaces},
pages = {383--394},
abstract = {We present a system and study of personalized energy-related recommendation. AgentSwitch utilizes electricity usage data collected from users' households over a period of time to realize a range of smart energy-related recommendations on energy tariffs, load detection and usage shifting. The web service is driven by a third party real-time energy tariff API (uSwitch), an energy data store, a set of algorithms for usage prediction, and appliance-level load disaggregation. We present the system design and user evaluation consisting of interviews and interface walkthroughs. We recruited participants from a previous study during which three months of their household's energy use was recorded to evaluate personalized recommendations in AgentSwitch. Our contributions are a) a systems architecture for personalized energy services; and b) findings from the evaluation that reveal challenges in designing energy-related recommender systems. In response to the challenges we formulate design recommendations to mitigate barriers to switching tariffs, to incentivize load shifting, and to automate energy management.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a system and study of personalized energy-related recommendation. AgentSwitch utilizes electricity usage data collected from users' households over a period of time to realize a range of smart energy-related recommendations on energy tariffs, load detection and usage shifting. The web service is driven by a third party real-time energy tariff API (uSwitch), an energy data store, a set of algorithms for usage prediction, and appliance-level load disaggregation. We present the system design and user evaluation consisting of interviews and interface walkthroughs. We recruited participants from a previous study during which three months of their household's energy use was recorded to evaluate personalized recommendations in AgentSwitch. Our contributions are a) a systems architecture for personalized energy services; and b) findings from the evaluation that reveal challenges in designing energy-related recommender systems. In response to the challenges we formulate design recommendations to mitigate barriers to switching tariffs, to incentivize load shifting, and to automate energy management.
Huynh, Trung Dong; Ebden, Mark; Venanzi, Matteo; Ramchurn, Sarvapali; Roberts, Stephen; Moreau, Luc
Interpretation of Crowdsourced Activities Using Provenance Network Analysis Proceedings Article
In: The First AAAI Conference on Human Computation and Crowdsourcing, Association for the Advancement of Artificial Intelligence, 2013.
@inproceedings{eps357199,
title = {Interpretation of Crowdsourced Activities Using Provenance Network Analysis},
author = {Trung Dong Huynh and Mark Ebden and Matteo Venanzi and Sarvapali Ramchurn and Stephen Roberts and Luc Moreau},
url = {http://eprints.soton.ac.uk/357199/},
year = {2013},
date = {2013-01-01},
booktitle = {The First AAAI Conference on Human Computation and Crowdsourcing},
publisher = {Association for the Advancement of Artificial Intelligence},
abstract = {Understanding the dynamics of a crowdsourcing application and controlling the quality of the data it generates is challenging, partly due to the lack of tools to do so. Provenance is a domain-independent means to represent what happened in an application, which can help verify data and infer their quality. It can also reveal the processes that led to a data item and the interactions of contributors with it. Provenance patterns can manifest real-world phenomena such as a significant interest in a piece of content, providing an indication of its quality, or even issues such as undesirable interactions within a group of contributors. This paper presents an application-independent methodology for analysing provenance graphs, constructed from provenance records, to learn about such patterns and to use them for assessing some key properties of crowdsourced data, such as their quality, in an automated manner. Validating this method on the provenance records of CollabMap, an online crowdsourcing mapping application, we demonstrated an accuracy level of over 95% for the trust classification of data generated by the crowd therein.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Understanding the dynamics of a crowdsourcing application and controlling the quality of the data it generates is challenging, partly due to the lack of tools to do so. Provenance is a domain-independent means to represent what happened in an application, which can help verify data and infer their quality. It can also reveal the processes that led to a data item and the interactions of contributors with it. Provenance patterns can manifest real-world phenomena such as a significant interest in a piece of content, providing an indication of its quality, or even issues such as undesirable interactions within a group of contributors. This paper presents an application-independent methodology for analysing provenance graphs, constructed from provenance records, to learn about such patterns and to use them for assessing some key properties of crowdsourced data, such as their quality, in an automated manner. Validating this method on the provenance records of CollabMap, an online crowdsourcing mapping application, we demonstrated an accuracy level of over 95% for the trust classification of data generated by the crowd therein.
Kleiner, Alexander; Farinelli, Alessandro; Ramchurn, Sarvapali; Shi, Bing; Mafioletti, Fabio; Refatto, Riccardo
RMASBench: a benchmarking system for multi-agent coordination in urban search and rescue Proceedings Article
In: International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS 2013), 2013.
@inproceedings{eps350678,
title = {RMASBench: a benchmarking system for multi-agent coordination in urban search and rescue},
author = {Alexander Kleiner and Alessandro Farinelli and Sarvapali Ramchurn and Bing Shi and Fabio Mafioletti and Riccardo Refatto},
url = {http://eprints.soton.ac.uk/350678/},
year = {2013},
date = {2013-01-01},
booktitle = {International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS 2013)},
abstract = {This demonstration paper illustrates RMASBench, a new benchmarking system based on the RoboCup Rescue Agent simulator. The aim of the system is to facilitate benchmarking of coordination approaches in controlled settings for dynamic rescue scenario. In particular, the key features of the systems are: i) programming interfaces to plug-in coordination algorithms without the need for implementing and tuning low-level agents? behaviors, ii) implementations of state-of-the art coordination approaches: DSA and MaxSum, iii) a large scale crowd simulator, which exploits GPUs parallel architecture, to simulate the behaviour of thousands of agents in real time.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This demonstration paper illustrates RMASBench, a new benchmarking system based on the RoboCup Rescue Agent simulator. The aim of the system is to facilitate benchmarking of coordination approaches in controlled settings for dynamic rescue scenario. In particular, the key features of the systems are: i) programming interfaces to plug-in coordination algorithms without the need for implementing and tuning low-level agents? behaviors, ii) implementations of state-of-the art coordination approaches: DSA and MaxSum, iii) a large scale crowd simulator, which exploits GPUs parallel architecture, to simulate the behaviour of thousands of agents in real time.
Ramchurn, Sarvapali; Osborne, Michael; Parson, Oliver; Rahwan, Talal; Maleki, Sasan; Reece, Steve; Huynh, Trung Dong; Alam, Muddasser; Fischer, Joel; Rodden, Tom; Moreau, Luc; Roberts, Sephen
AgentSwitch: towards smart electricity tariff selection Proceedings Article
In: 12th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2013), International Foundation for Autonomous Agents and Multiagent Systems, 2013.
@inproceedings{eps349815,
title = {AgentSwitch: towards smart electricity tariff selection},
author = {Sarvapali Ramchurn and Michael Osborne and Oliver Parson and Talal Rahwan and Sasan Maleki and Steve Reece and Trung Dong Huynh and Muddasser Alam and Joel Fischer and Tom Rodden and Luc Moreau and Sephen Roberts},
url = {http://eprints.soton.ac.uk/349815/},
year = {2013},
date = {2013-01-01},
booktitle = {12th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2013)},
publisher = {International Foundation for Autonomous Agents and Multiagent Systems},
abstract = {In this paper, we present AgentSwitch, a prototype agent-based platform to solve the electricity tariff selection problem. AgentSwitch incorporates novel algorithms to make predictions of hourly energy usage as well as detect (and suggest to the user) deferrable loads that could be shifted to off-peak times to maximise savings. To take advantage of group discounts from energy retailers, we develop a new scalable collective energy purchasing mechanism, based on the Shapley value, that ensures individual members of a collective (interacting through AgentSwitch) fairly share the discounts. To demonstrate the effectiveness of our algorithms we empirically evaluate them individually on real-world data (with up to 3000 homes in the UK) and show that they outperform the state of the art in their domains. Finally, to ensure individual components are accountable in providing recommendations, we provide a novel provenance-tracking service to record the ?ow of data in the system, and therefore provide users with a means of checking the provenance of suggestions from AgentSwitch and assess their reliability.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we present AgentSwitch, a prototype agent-based platform to solve the electricity tariff selection problem. AgentSwitch incorporates novel algorithms to make predictions of hourly energy usage as well as detect (and suggest to the user) deferrable loads that could be shifted to off-peak times to maximise savings. To take advantage of group discounts from energy retailers, we develop a new scalable collective energy purchasing mechanism, based on the Shapley value, that ensures individual members of a collective (interacting through AgentSwitch) fairly share the discounts. To demonstrate the effectiveness of our algorithms we empirically evaluate them individually on real-world data (with up to 3000 homes in the UK) and show that they outperform the state of the art in their domains. Finally, to ensure individual components are accountable in providing recommendations, we provide a novel provenance-tracking service to record the ?ow of data in the system, and therefore provide users with a means of checking the provenance of suggestions from AgentSwitch and assess their reliability.
Ramchurn, Sarvapali D.; Huynh, Trung Dong; Venanzi, Matteo; Shi, Bing
Collabmap: crowdsourcing maps for emergency planning Proceedings Article
In: The 5th Annual ACM Web Science Conference, pp. 326–335, 2013.
@inproceedings{eps350677,
title = {Collabmap: crowdsourcing maps for emergency planning},
author = {Sarvapali D. Ramchurn and Trung Dong Huynh and Matteo Venanzi and Bing Shi},
url = {http://eprints.soton.ac.uk/350677/},
year = {2013},
date = {2013-01-01},
booktitle = {The 5th Annual ACM Web Science Conference},
pages = {326--335},
abstract = {In this paper, we present a software tool to help emergency planners at Hampshire County Council in the UK to create maps for high-fidelity crowd simulations that require evacuation routes from buildings to roads. The main feature of the system is a crowdsourcing mechanism that breaks down the problem of creating evacuation routes into microtasks that a contributor to the platform can execute in less than a minute. As part of the mechanism we developed a concensus-based trust mechanism that filters out incorrect contributions and ensures that the individual tasks are complete and correct. To drive people to contribute to the platform, we experimented with different incentive mechanisms and applied these over different time scales, the aim being to evaluate what incentives work with different types of crowds, including anonymous contributors from Amazon Mechanical Turk. The results of the 'in the wild' deployment of the system show that the system is effective at engaging contributors to perform tasks correctly and that users respond to incentives in different ways. More specifically, we show that purely social motives are not good enough to attract a large number of contributors and that contributors are averse to the uncertainty in winning rewards. When taken altogether, our results suggest that a combination of incentives may be the best approach to harnessing the maximum number of resources to get socially valuable tasks (such for planning applications) performed on a large scale.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we present a software tool to help emergency planners at Hampshire County Council in the UK to create maps for high-fidelity crowd simulations that require evacuation routes from buildings to roads. The main feature of the system is a crowdsourcing mechanism that breaks down the problem of creating evacuation routes into microtasks that a contributor to the platform can execute in less than a minute. As part of the mechanism we developed a concensus-based trust mechanism that filters out incorrect contributions and ensures that the individual tasks are complete and correct. To drive people to contribute to the platform, we experimented with different incentive mechanisms and applied these over different time scales, the aim being to evaluate what incentives work with different types of crowds, including anonymous contributors from Amazon Mechanical Turk. The results of the 'in the wild' deployment of the system show that the system is effective at engaging contributors to perform tasks correctly and that users respond to incentives in different ways. More specifically, we show that purely social motives are not good enough to attract a large number of contributors and that contributors are averse to the uncertainty in winning rewards. When taken altogether, our results suggest that a combination of incentives may be the best approach to harnessing the maximum number of resources to get socially valuable tasks (such for planning applications) performed on a large scale.
Rigas, Emmanouil; Ramchurn, Sarvapali; Bassiliades, Nick; Koutitas, Georgios
Congestion management for urban EV charging systems Proceedings Article
In: 4th IEEE International Conference on Smart Grid Communications (SmartGridComm), IEEE, 2013.
@inproceedings{eps356081,
title = {Congestion management for urban EV charging systems},
author = {Emmanouil Rigas and Sarvapali Ramchurn and Nick Bassiliades and Georgios Koutitas},
url = {http://eprints.soton.ac.uk/356081/},
year = {2013},
date = {2013-01-01},
booktitle = {4th IEEE International Conference on Smart Grid Communications (SmartGridComm)},
volume = {4},
publisher = {IEEE},
abstract = {We consider the problem of managing Electric Vehicle (EV) charging at charging points in a city to ensure that the load on the charging points remains within the desired limits while minimizing the inconvenience to EV owners. We develop solutions that treat charging points and EV users as self-interested agents that aim to maximize their profit and minimize the impact on their schedule. In particular, we propose variants of a decentralised and dynamic approach as well as an optimal centralised static approach. We evaluated these solutions in a real setting based on the road network and the location of parking garages of a UK city and show that the optimal centralised (non-dynamic) solution manages the congestion the best but does not scale well, while the decentralised solutions scale to thousands of agents.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We consider the problem of managing Electric Vehicle (EV) charging at charging points in a city to ensure that the load on the charging points remains within the desired limits while minimizing the inconvenience to EV owners. We develop solutions that treat charging points and EV users as self-interested agents that aim to maximize their profit and minimize the impact on their schedule. In particular, we propose variants of a decentralised and dynamic approach as well as an optimal centralised static approach. We evaluated these solutions in a real setting based on the road network and the location of parking garages of a UK city and show that the optimal centralised (non-dynamic) solution manages the congestion the best but does not scale well, while the decentralised solutions scale to thousands of agents.
Svensson, Kim; Ramchurn, Sarvapali; Cruz, Francisco; Rodriguez-Aguilar, Juan-Antonio; Cerquides, Jesus
Solving the coalition structure generation problem on a GPU Proceedings Article
In: 6th International Workshop on Optimisation in Multi-Agent Systems, 2013.
@inproceedings{eps352204,
title = {Solving the coalition structure generation problem on a GPU},
author = {Kim Svensson and Sarvapali Ramchurn and Francisco Cruz and Juan-Antonio Rodriguez-Aguilar and Jesus Cerquides},
url = {http://eprints.soton.ac.uk/352204/},
year = {2013},
date = {2013-01-01},
booktitle = {6th International Workshop on Optimisation in Multi-Agent Systems},
abstract = {We develop the first parallel algorithm for Coalition Structure Generation (CSG), which is central to many multi-agent systems applications. Our approach involves distributing the key steps of a dynamic programming approach to CSG across computational nodes on a Graphics Processing Unit (GPU) such that each of the thousands of threads of computation can be used to perform small computations that speed up the overall process. In so doing, we solve important challenges that arise in solving combinatorial optimisation problems on GPUs such as the efficient allocation of memory and computational threads to every step of the algorithm. In our empirical evaluations on a standard GPU, our results show an improvement of orders of magnitude over current dynamic programming approaches with an ever increasing divergence between the CPU and GPU-based algorithms in terms of growth. Thus, our algorithm is able to solve the CSG problem for 29 agents in one hour and thirty minutes as opposed to three days for the current state of the art dynamic programming algorithms.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We develop the first parallel algorithm for Coalition Structure Generation (CSG), which is central to many multi-agent systems applications. Our approach involves distributing the key steps of a dynamic programming approach to CSG across computational nodes on a Graphics Processing Unit (GPU) such that each of the thousands of threads of computation can be used to perform small computations that speed up the overall process. In so doing, we solve important challenges that arise in solving combinatorial optimisation problems on GPUs such as the efficient allocation of memory and computational threads to every step of the algorithm. In our empirical evaluations on a standard GPU, our results show an improvement of orders of magnitude over current dynamic programming approaches with an ever increasing divergence between the CPU and GPU-based algorithms in terms of growth. Thus, our algorithm is able to solve the CSG problem for 29 agents in one hour and thirty minutes as opposed to three days for the current state of the art dynamic programming algorithms.
Truong, Ngoc Cuong; McInerney, James; Tran-Thanh, Long; Costanza, Enrico; Ramchurn, Sarvapali D.
Forecasting multi-appliance usage for smart home energy management Proceedings Article
In: 23rd International Joint Conference on Artificial Intelligence (IJCAI 2013), 2013.
@inproceedings{eps351242,
title = {Forecasting multi-appliance usage for smart home energy management},
author = {Ngoc Cuong Truong and James McInerney and Long Tran-Thanh and Enrico Costanza and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/351242/},
year = {2013},
date = {2013-01-01},
booktitle = {23rd International Joint Conference on Artificial Intelligence (IJCAI 2013)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Truong, Ngoc Cuong; Tran-Thanh, Long; Costanza, Enrico; Ramchurn, D. Sarvapali
Activity prediction for agent-based home energy management Proceedings Article
In: Agent Technologies for Energy Systems (ATES 2013), 2013.
@inproceedings{eps351238,
title = {Activity prediction for agent-based home energy management},
author = {Ngoc Cuong Truong and Long Tran-Thanh and Enrico Costanza and D. Sarvapali Ramchurn},
url = {http://eprints.soton.ac.uk/351238/},
year = {2013},
date = {2013-01-01},
booktitle = {Agent Technologies for Energy Systems (ATES 2013)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Truong, Ngoc Cuong; Tran-Thanh, Long; Costanza, Enrico; Ramchurn, Sarvapali D.
Towards appliance usage prediction for home energy management Proceedings Article
In: ACM E-Energy 2013, 2013.
@inproceedings{eps351240,
title = {Towards appliance usage prediction for home energy management},
author = {Ngoc Cuong Truong and Long Tran-Thanh and Enrico Costanza and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/351240/},
year = {2013},
date = {2013-01-01},
booktitle = {ACM E-Energy 2013},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2012
Costanza, Enrico; Ramchurn, Sarvapali D.; Jennings, Nicholas R.
Understanding domestic energy consumption through interactive visualisation: a field study Proceedings Article
In: UbiComp '12. Proceedings of the 2012 ACM Conference on Ubiquitous Computing, pp. 216–225, 2012.
@inproceedings{eps338804,
title = {Understanding domestic energy consumption through interactive visualisation: a field study},
author = {Enrico Costanza and Sarvapali D. Ramchurn and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/338804/},
year = {2012},
date = {2012-01-01},
booktitle = {UbiComp '12. Proceedings of the 2012 ACM Conference on Ubiquitous Computing},
pages = {216--225},
abstract = {Motivated by the need to better manage energy demand in the home, in this paper we advocate the integration into Ubicomp systems of interactive energy consumption visualisations, that allow users to engage with and understand their consumption data, relating it to concrete activities in their life. To this end, we present the design, implementation, and evaluation of FigureEnergy, a novel interactive visualisation that allows users to annotate and manipulate a graphical representation of their own electricity consumption data, and therefore make sense of their past energy usage and understand when, how, and to what end, some amount of energy was used. To validate our design, we deployed FigureEnergy ?in the wild? ? 12 participants installed meters in their homes and used the system for a period of two weeks. The results suggest that the annotation approach is successful overall: by engaging with the data users started to relate energy consumption to activities rather than just to appliances. Moreover, they were able to discover that some appliances consume more than they expected, despite having had prior experience of using other electricity displays.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Motivated by the need to better manage energy demand in the home, in this paper we advocate the integration into Ubicomp systems of interactive energy consumption visualisations, that allow users to engage with and understand their consumption data, relating it to concrete activities in their life. To this end, we present the design, implementation, and evaluation of FigureEnergy, a novel interactive visualisation that allows users to annotate and manipulate a graphical representation of their own electricity consumption data, and therefore make sense of their past energy usage and understand when, how, and to what end, some amount of energy was used. To validate our design, we deployed FigureEnergy ?in the wild? ? 12 participants installed meters in their homes and used the system for a period of two weeks. The results suggest that the annotation approach is successful overall: by engaging with the data users started to relate energy consumption to activities rather than just to appliances. Moreover, they were able to discover that some appliances consume more than they expected, despite having had prior experience of using other electricity displays.
Ebden, Mark; Huynh, Trung Dong; Moreau, Luc; Ramchurn, Sarvapali; Stephen, Roberts
Network analysis on provenance graphs from a crowdsourcing application Proceedings Article
In: Groth, Paul; Frew, James (Ed.): 4th International Provenance and Annotation Workshop, pp. 168–182, 2012.
@inproceedings{eps340068,
title = {Network analysis on provenance graphs from a crowdsourcing application},
author = {Mark Ebden and Trung Dong Huynh and Luc Moreau and Sarvapali Ramchurn and Roberts Stephen},
editor = {Paul Groth and James Frew},
url = {http://eprints.soton.ac.uk/340068/},
year = {2012},
date = {2012-01-01},
booktitle = {4th International Provenance and Annotation Workshop},
volume = {7525},
pages = {168--182},
series = {0302-9743},
abstract = {Crowdsourcing has become a popular means for quickly achieving various tasks in large quantities. CollabMap is an online mapping application in which we crowdsource the identification of evacuation routes in residential areas to be used for planning large-scale evacuations. So far, approximately 38,000 micro-tasks have been completed by over 100 contributors. In order to assist with data verification, we introduced provenance tracking into the application, and approximately 5,000 provenance graphs have been generated. They have provided us various insights into the typical characteristics of provenance graphs in the crowdsourcing context. In particular, we have estimated probability distribution functions over three selected characteristics of these provenance graphs: the node degree, the graph diameter, and the densification exponent. We describe methods to define these three characteristics across specific combinations of node types and edge types, and present our findings in this paper. Applications of our methods include rapid comparison of one provenance graph versus another, or of one style of provenance database versus another. Our results also indicate that provenance graphs represent a suitable area of exploitation for existing network analysis tools concerned with modelling, prediction, and the inference of missing nodes and edges.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Crowdsourcing has become a popular means for quickly achieving various tasks in large quantities. CollabMap is an online mapping application in which we crowdsource the identification of evacuation routes in residential areas to be used for planning large-scale evacuations. So far, approximately 38,000 micro-tasks have been completed by over 100 contributors. In order to assist with data verification, we introduced provenance tracking into the application, and approximately 5,000 provenance graphs have been generated. They have provided us various insights into the typical characteristics of provenance graphs in the crowdsourcing context. In particular, we have estimated probability distribution functions over three selected characteristics of these provenance graphs: the node degree, the graph diameter, and the densification exponent. We describe methods to define these three characteristics across specific combinations of node types and edge types, and present our findings in this paper. Applications of our methods include rapid comparison of one provenance graph versus another, or of one style of provenance database versus another. Our results also indicate that provenance graphs represent a suitable area of exploitation for existing network analysis tools concerned with modelling, prediction, and the inference of missing nodes and edges.
Matthews, Tim; Ramchurn, Sarvapali; Chalkiadakis, Georgios
Competing with humans at fantasy football: team formation in large partially-observable domains Proceedings Article
In: Proceedings of the Twenty-Sixth Conference on Artificial Intelligence, pp. 1394–1400, Association for the Advancement of Artificial Intelligence, 2012.
@inproceedings{eps340382,
title = {Competing with humans at fantasy football: team formation in large partially-observable domains},
author = {Tim Matthews and Sarvapali Ramchurn and Georgios Chalkiadakis},
url = {http://eprints.soton.ac.uk/340382/},
year = {2012},
date = {2012-01-01},
booktitle = {Proceedings of the Twenty-Sixth Conference on Artificial Intelligence},
pages = {1394--1400},
publisher = {Association for the Advancement of Artificial Intelligence},
abstract = {We present the first real-world benchmark for sequentially optimal team formation, working within the framework of a class of online football prediction games known as Fantasy Football. We model the problem as a Bayesian reinforcement learning one, where the action space is exponential in the number of players and where the decision maker?s beliefs are over multiple characteristics of each footballer. We then exploit domain knowledge to construct computationally tractable solution techniques in order to build a competitive automated Fantasy Football manager. Thus, we are able to establish the baseline performance in this domain, even without complete information on footballers? performances (accessible to human managers), showing that our agent is able to rank at around the top percentile when pitched against 2.5M human players},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present the first real-world benchmark for sequentially optimal team formation, working within the framework of a class of online football prediction games known as Fantasy Football. We model the problem as a Bayesian reinforcement learning one, where the action space is exponential in the number of players and where the decision maker?s beliefs are over multiple characteristics of each footballer. We then exploit domain knowledge to construct computationally tractable solution techniques in order to build a competitive automated Fantasy Football manager. Thus, we are able to establish the baseline performance in this domain, even without complete information on footballers? performances (accessible to human managers), showing that our agent is able to rank at around the top percentile when pitched against 2.5M human players
Miller, Sam; Ramchurn, Sarvapali D; Rogers, Alex
Optimal Decentralised Dispatch of Embedded Generation in the Smart Grid Journal Article
In: In Proc. 11th Int. Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), 2012.
@article{eps273142,
title = {Optimal Decentralised Dispatch of Embedded Generation in the Smart Grid},
author = {Sam Miller and Sarvapali D Ramchurn and Alex Rogers},
url = {http://eprints.soton.ac.uk/273142/},
year = {2012},
date = {2012-01-01},
booktitle = {Proc. 11th Int. Conference on Autonomous Agents and Multi-Agent Systems (AAMAS)},
journal = {In Proc. 11th Int. Conference on Autonomous Agents and Multi-Agent Systems (AAMAS)},
abstract = {Distribution network operators face a number of challenges; capacity constrained networks, and balancing electricity demand with generation from intermittent renewable resources. Thus, there is an increasing need for scalable approaches to facilitate optimal dispatch in the distribution network. To this end, we cast the optimal dispatch problem as a decentralised agent-based coordination problem and formalise it as a DCOP. We show how this can be decomposed as a factor graph and solved in a decentralised manner using algorithms based on the generalised distributive law; in particular, the max-sum algorithm. We go on to show that max-sum applied na?vely in this setting performs a large number of redundant computations. To address this issue, we present a novel decentralised message passing algorithm using dynamic programming that outperforms max-sum by pruning the search space. We empirically evaluate our algorithm using real distribution network data, showing that it outperforms (in terms of computational time and total size of messages sent) both a centralised approach, which uses IBM?s ILOG CPLEX 12.2, and max-sum, for large networks.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Distribution network operators face a number of challenges; capacity constrained networks, and balancing electricity demand with generation from intermittent renewable resources. Thus, there is an increasing need for scalable approaches to facilitate optimal dispatch in the distribution network. To this end, we cast the optimal dispatch problem as a decentralised agent-based coordination problem and formalise it as a DCOP. We show how this can be decomposed as a factor graph and solved in a decentralised manner using algorithms based on the generalised distributive law; in particular, the max-sum algorithm. We go on to show that max-sum applied na?vely in this setting performs a large number of redundant computations. To address this issue, we present a novel decentralised message passing algorithm using dynamic programming that outperforms max-sum by pruning the search space. We empirically evaluate our algorithm using real distribution network data, showing that it outperforms (in terms of computational time and total size of messages sent) both a centralised approach, which uses IBM?s ILOG CPLEX 12.2, and max-sum, for large networks.
Ramchurn, Sarvapali; Vytelingum, Perukrishnen; Rogers, Alex; Jennings, Nicholas R.
Putting the Smarts into the Smart Grid: A Grand Challenge for Artificial Intelligence Journal Article
In: Communications of the ACM, vol. 55, no. 4, pp. 86–97, 2012.
@article{eps272606,
title = {Putting the Smarts into the Smart Grid: A Grand Challenge for Artificial Intelligence},
author = {Sarvapali Ramchurn and Perukrishnen Vytelingum and Alex Rogers and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/272606/},
year = {2012},
date = {2012-01-01},
journal = {Communications of the ACM},
volume = {55},
number = {4},
pages = {86--97},
publisher = {ACM},
abstract = {The phenomenal growth in material wealth experienced in developed countries throughout the twentieth century has largely been driven by the availability of cheap energy derived from fossil fuels (originally coal, then oil, and most recently natural gas). However, the continued availability of this cheap energy cannot be taken for granted given the growing concern that increasing demand for these fuels (and particularly, demand for oil) will outstrip our ability to produce them (so called `peak oil'). Many mature oil and gas fields around the world have already peaked and their annual production is now steadily declining. Predictions of when world oil production will peak vary between 0-20 years into the future, but even the most conservative estimates provide little scope for complacency given the significant price increases that peak oil is likely to precipitate. Furthermore, many of the oil and gas reserves that do remain are in environmentally or politically sensitive regions of the world where threats to supply create increased price volatility (as evidenced by the 2010 Deepwater Horizon disaster and 2011 civil unrest in the Middle East). Finally, the growing consensus on the long term impact of carbon emissions from burning fossil fuels suggests that even if peak oil is avoided, and energy security assured, a future based on fossil fuel use will expose regions of the world to damaging climate change that will make the lives of many of the world's poorest people even harder.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The phenomenal growth in material wealth experienced in developed countries throughout the twentieth century has largely been driven by the availability of cheap energy derived from fossil fuels (originally coal, then oil, and most recently natural gas). However, the continued availability of this cheap energy cannot be taken for granted given the growing concern that increasing demand for these fuels (and particularly, demand for oil) will outstrip our ability to produce them (so called `peak oil'). Many mature oil and gas fields around the world have already peaked and their annual production is now steadily declining. Predictions of when world oil production will peak vary between 0-20 years into the future, but even the most conservative estimates provide little scope for complacency given the significant price increases that peak oil is likely to precipitate. Furthermore, many of the oil and gas reserves that do remain are in environmentally or politically sensitive regions of the world where threats to supply create increased price volatility (as evidenced by the 2010 Deepwater Horizon disaster and 2011 civil unrest in the Middle East). Finally, the growing consensus on the long term impact of carbon emissions from burning fossil fuels suggests that even if peak oil is avoided, and energy security assured, a future based on fossil fuel use will expose regions of the world to damaging climate change that will make the lives of many of the world's poorest people even harder.
Ramchurn, Sarvapali D.; Gerding, Enrico; Jennings, N. R.; Hu, Jun
Practical distributed coalition formation via heuristic negotiation in social networks Proceedings Article
In: Fifth International Workshop on Optimisation in Multi-Agent Systems (OPTMAS), 2012.
@inproceedings{eps344492,
title = {Practical distributed coalition formation via heuristic negotiation in social networks},
author = {Sarvapali D. Ramchurn and Enrico Gerding and N.R. Jennings and Jun Hu},
url = {http://eprints.soton.ac.uk/344492/},
year = {2012},
date = {2012-01-01},
booktitle = {Fifth International Workshop on Optimisation in Multi-Agent Systems (OPTMAS)},
abstract = {We present a novel framework for decentralised coalition formation in social networks, where agents can form coalitions through bilateral negotiations with their neighbours. Specifically, we present a practical negotiation protocol and decision functions that enable agents to form coalitions with agents beyond their peers. Building on this, we establish baseline negotiation strategies which we empirically show to be efficient (agreements are reached in few negotiation rounds) and effective (agreements have high utility compared to a centralised approach) on a variety of network topologies. Moreover, we show that the average degree of social networks can significantly affect the performance of these strategies.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a novel framework for decentralised coalition formation in social networks, where agents can form coalitions through bilateral negotiations with their neighbours. Specifically, we present a practical negotiation protocol and decision functions that enable agents to form coalitions with agents beyond their peers. Building on this, we establish baseline negotiation strategies which we empirically show to be efficient (agreements are reached in few negotiation rounds) and effective (agreements have high utility compared to a centralised approach) on a variety of network topologies. Moreover, we show that the average degree of social networks can significantly affect the performance of these strategies.
Richardson, Darren P.; Costanza, Enrico; Ramchurn, Sarvapali D.
Evaluating semi-automatic annotation of domestic energy consumption as a memory aid Proceedings Article
In: UbiComp '12 Proceedings of the 2012 ACM Conference on Ubiquitous Computing, pp. 613–614, 2012.
@inproceedings{eps349083,
title = {Evaluating semi-automatic annotation of domestic energy consumption as a memory aid},
author = {Darren P. Richardson and Enrico Costanza and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/349083/},
year = {2012},
date = {2012-01-01},
booktitle = {UbiComp '12 Proceedings of the 2012 ACM Conference on Ubiquitous Computing},
pages = {613--614},
abstract = {Frequent feedback about energy consumption can help conservation, one of the current global challenges. Such feedback is most helpful if users can relate it to their own day-to-day activities. In earlier work we showed that manual annotation of domestic energy consumption logs aids users to make such connection and discover patterns they were not aware of. In this poster we report how we augmented manual annotation with machine learning classification techniques. We propose the design of a lab study to evaluate the system, extending methods used to evaluate context aware memory aids, and we present the results of a pilot with 5 participants.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Frequent feedback about energy consumption can help conservation, one of the current global challenges. Such feedback is most helpful if users can relate it to their own day-to-day activities. In earlier work we showed that manual annotation of domestic energy consumption logs aids users to make such connection and discover patterns they were not aware of. In this poster we report how we augmented manual annotation with machine learning classification techniques. We propose the design of a lab study to evaluate the system, extending methods used to evaluate context aware memory aids, and we present the results of a pilot with 5 participants.
Rogers, Alex; Ramchurn, Sarvapali; Jennings, Nicholas R.
Delivering the smart grid: Challenges for autonomous agents and multi-agent systems research Proceedings Article
In: Twenty-Sixth AAAI Conference on Artificial Intelligence (AAAI-12), pp. 2166–2172, 2012.
@inproceedings{eps337560,
title = {Delivering the smart grid: Challenges for autonomous agents and multi-agent systems research},
author = {Alex Rogers and Sarvapali Ramchurn and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/337560/},
year = {2012},
date = {2012-01-01},
booktitle = {Twenty-Sixth AAAI Conference on Artificial Intelligence (AAAI-12)},
pages = {2166--2172},
abstract = {Restructuring electricity grids to meet the increased demand caused by the electrification of transport and heating, while making greater use of intermittent renewable energy sources, represents one of the greatest engineering challenges of our day. This modern electric- ity grid, in which both electricity and information flow in two directions between large numbers of widely dis- tributed suppliers and generators -- commonly termed the ?smart grid? -- represents a radical reengineering of infrastructure which has changed little over the last hundred years. However, the autonomous behaviour expected of the smart grid, its distributed nature, and the existence of multiple stakeholders each with their own incentives and interests, challenges existing engineering approaches. In this challenge paper, we describe why we believe that artificial intelligence, and particularly, the fields of autonomous agents and multi-agent systems are essential for delivering the smart grid as it is envisioned. We present some recent work in this area and describe many of the challenges that still remain.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Restructuring electricity grids to meet the increased demand caused by the electrification of transport and heating, while making greater use of intermittent renewable energy sources, represents one of the greatest engineering challenges of our day. This modern electric- ity grid, in which both electricity and information flow in two directions between large numbers of widely dis- tributed suppliers and generators -- commonly termed the ?smart grid? -- represents a radical reengineering of infrastructure which has changed little over the last hundred years. However, the autonomous behaviour expected of the smart grid, its distributed nature, and the existence of multiple stakeholders each with their own incentives and interests, challenges existing engineering approaches. In this challenge paper, we describe why we believe that artificial intelligence, and particularly, the fields of autonomous agents and multi-agent systems are essential for delivering the smart grid as it is envisioned. We present some recent work in this area and describe many of the challenges that still remain.
Truong, Ngoc Cuong; Tran-Thanh, Long; Costanza, Enrico; Ramchurn, Sarvapali D.
Predicting energy consumption activities for home energy management Proceedings Article
In: Agent Technologies for Energy Systems (ATES 2012), 2012.
@inproceedings{eps339215,
title = {Predicting energy consumption activities for home energy management},
author = {Ngoc Cuong Truong and Long Tran-Thanh and Enrico Costanza and Sarvapali D. Ramchurn},
url = {http://eprints.soton.ac.uk/339215/},
year = {2012},
date = {2012-01-01},
booktitle = {Agent Technologies for Energy Systems (ATES 2012)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Voice, Thomas; Ramchurn, Sarvapali; Jennings, Nick
On coalition formation with sparse synergies Proceedings Article
In: Proc. 11th Int. Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), pp. 223–230, 2012.
@inproceedings{eps273083,
title = {On coalition formation with sparse synergies},
author = {Thomas Voice and Sarvapali Ramchurn and Nick Jennings},
url = {http://eprints.soton.ac.uk/273083/},
year = {2012},
date = {2012-01-01},
booktitle = {Proc. 11th Int. Conference on Autonomous Agents and Multi-Agent Systems (AAMAS)},
pages = {223--230},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Simpson, Edwin; Reece, Steven; Penta, Antonio; Ramchurn, Sarvapali D
Using a Bayesian Model to Combine LDA Features with Crowdsourced Responses Proceedings Article
In: Proceedings of The Twenty-First Text REtrieval Conference, TREC 2012, Gaithersburg, Maryland, USA, November 6-9, 2012, 2012.
@inproceedings{DBLP:conf/trec/SimpsonRPR12,
title = {Using a Bayesian Model to Combine LDA Features with Crowdsourced
Responses},
author = {Edwin Simpson and Steven Reece and Antonio Penta and Sarvapali D Ramchurn},
url = {http://trec.nist.gov/pubs/trec21/papers/HAC.crowd.final.pdf},
year = {2012},
date = {2012-01-01},
booktitle = {Proceedings of The Twenty-First Text REtrieval Conference, TREC
2012, Gaithersburg, Maryland, USA, November 6-9, 2012},
crossref = {DBLP:conf/trec/2012},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2011
Alam, Muddasser; Rogers, Alex; Ramchurn, Sarvapali
A negotiation protocol for multiple interdependent issues negotiation over energy exchange Proceedings Article
In: IJCAI Workshop on AI for an Intelligent Planet, 2011, (Event Dates: July-16).
@inproceedings{eps272479,
title = {A negotiation protocol for multiple interdependent issues negotiation over energy exchange},
author = {Muddasser Alam and Alex Rogers and Sarvapali Ramchurn},
url = {http://eprints.soton.ac.uk/272479/},
year = {2011},
date = {2011-01-01},
booktitle = {IJCAI Workshop on AI for an Intelligent Planet},
abstract = {We present a novel negotiation protocol to facilitate energy exchange between off-grid homes that are equipped with renewable energy generation and electricity storage. Our solution imposes additional constraints on negotiation such that it reduces a complex interdependent multi-issue problem to one that is tractable. We prove that using our protocol, agents can reach a Pareto-optimal, dominant strategy equilibrium in a decentralized and timely fashion. We empirically evaluate our approach in a realistic setting. In this case, we show that energy exchange can be useful in reducing the capacity of the energy storage devices in homes by close to 40%},
note = {Event Dates: July-16},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a novel negotiation protocol to facilitate energy exchange between off-grid homes that are equipped with renewable energy generation and electricity storage. Our solution imposes additional constraints on negotiation such that it reduces a complex interdependent multi-issue problem to one that is tractable. We prove that using our protocol, agents can reach a Pareto-optimal, dominant strategy equilibrium in a decentralized and timely fashion. We empirically evaluate our approach in a realistic setting. In this case, we show that energy exchange can be useful in reducing the capacity of the energy storage devices in homes by close to 40%
Macarthur, Kathryn; Stranders, Ruben; Ramchurn, Sarvapali; Jennings, Nick
A Distributed Anytime Algorithm for Dynamic Task Allocation in Multi-Agent Systems Proceedings Article
In: Twenty-Fifth Conference on Artificial Intelligence (AAAI), pp. 701–706, AAAI Press, 2011, (Event Dates: August 7-11, 2011).
@inproceedings{eps272233,
title = {A Distributed Anytime Algorithm for Dynamic Task Allocation in Multi-Agent Systems},
author = {Kathryn Macarthur and Ruben Stranders and Sarvapali Ramchurn and Nick Jennings},
url = {http://eprints.soton.ac.uk/272233/},
year = {2011},
date = {2011-01-01},
booktitle = {Twenty-Fifth Conference on Artificial Intelligence (AAAI)},
pages = {701--706},
publisher = {AAAI Press},
abstract = {We introduce a novel distributed algorithm for multi-agent task allocation problems where the sets of tasks and agents constantly change over time. We build on an existing anytime algorithm (fast-max-sum), and give it significant new capa- bilities: namely, an online pruning procedure that simplifies the problem, and a branch-and-bound technique that reduces the search space. This allows us to scale to problems with hundreds of tasks and agents. We empirically evaluate our algorithm against established benchmarks and find that, even in such large environments, a solution is found up to 31% faster, and with up to 23% more utility, than state-of-the-art approximation algorithms. In addition, our algorithm sends up to 30% fewer messages than current approaches when the set of agents or tasks changes.},
note = {Event Dates: August 7-11, 2011},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce a novel distributed algorithm for multi-agent task allocation problems where the sets of tasks and agents constantly change over time. We build on an existing anytime algorithm (fast-max-sum), and give it significant new capa- bilities: namely, an online pruning procedure that simplifies the problem, and a branch-and-bound technique that reduces the search space. This allows us to scale to problems with hundreds of tasks and agents. We empirically evaluate our algorithm against established benchmarks and find that, even in such large environments, a solution is found up to 31% faster, and with up to 23% more utility, than state-of-the-art approximation algorithms. In addition, our algorithm sends up to 30% fewer messages than current approaches when the set of agents or tasks changes.
Macarthur, Kathryn; Vinyals, Meritxell; Farinelli, Alessandro; Ramchurn, Sarvapali; Jennings, Nick
Decentralised Parallel Machine Scheduling for Multi-Agent Task Allocation Proceedings Article
In: Fourth International Workshop on Optimisation in Multi-Agent Systems, 2011, (Event Dates: May 3, 2011).
@inproceedings{eps272234,
title = {Decentralised Parallel Machine Scheduling for Multi-Agent Task Allocation},
author = {Kathryn Macarthur and Meritxell Vinyals and Alessandro Farinelli and Sarvapali Ramchurn and Nick Jennings},
url = {http://eprints.soton.ac.uk/272234/},
year = {2011},
date = {2011-01-01},
booktitle = {Fourth International Workshop on Optimisation in Multi-Agent Systems},
abstract = {Multi-agent task allocation problems pervade a wide range of real-world applications, such as search and rescue in disaster manage- ment, or grid computing. In these applications, where agents are given tasks to perform in parallel, it is often the case that the performance of all agents is judged based on the time taken by the slowest agent to complete its tasks. Hence, efficient distribution of tasks across het- erogeneous agents is important to ensure a short completion time. An equivalent problem to this can be found in operations research, and is known as scheduling jobs on unrelated parallel machines (also known as Rensuremath|ensuremath|Cmax). In this paper, we draw parallels between unrelated parallel machine scheduling and multi-agent task allocation problems, and, in so doing, we present the decentralised task distribution algorithm (DTDA), the first decentralised solution to Rensuremath|ensuremath|Cmax. Empirical evaluation of the DTDA is shown to generate solutions within 86?97% of the optimal on sparse graphs, in the best case, whilst providing a very good estimate (within 1%) of the global solution at each agent.},
note = {Event Dates: May 3, 2011},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Multi-agent task allocation problems pervade a wide range of real-world applications, such as search and rescue in disaster manage- ment, or grid computing. In these applications, where agents are given tasks to perform in parallel, it is often the case that the performance of all agents is judged based on the time taken by the slowest agent to complete its tasks. Hence, efficient distribution of tasks across het- erogeneous agents is important to ensure a short completion time. An equivalent problem to this can be found in operations research, and is known as scheduling jobs on unrelated parallel machines (also known as Rensuremath|ensuremath|Cmax). In this paper, we draw parallels between unrelated parallel machine scheduling and multi-agent task allocation problems, and, in so doing, we present the decentralised task distribution algorithm (DTDA), the first decentralised solution to Rensuremath|ensuremath|Cmax. Empirical evaluation of the DTDA is shown to generate solutions within 86?97% of the optimal on sparse graphs, in the best case, whilst providing a very good estimate (within 1%) of the global solution at each agent.
Osborne, Michael A.; Rogers, Alex; Roberts, Stephen J.; Ramchurn, Sarvapali D.; Jennings, Nicholas R.
Gaussian Processes for Time Series Prediction Book Section
In: Bayesian Time Series Models, pp. 341–360, Cambridge University Press, 2011, (Chapter: 16).
@incollection{eps272746,
title = {Gaussian Processes for Time Series Prediction},
author = {Michael A. Osborne and Alex Rogers and Stephen J. Roberts and Sarvapali D. Ramchurn and Nicholas R. Jennings},
url = {http://eprints.soton.ac.uk/272746/},
year = {2011},
date = {2011-01-01},
booktitle = {Bayesian Time Series Models},
pages = {341--360},
publisher = {Cambridge University Press},
note = {Chapter: 16},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Ramchurn, Sarvapali; Vytelingum, Perukrishnen; Rogers, Alex; Jennings, Nick
Agent-based homeostatic control for green energy in the smart grid Journal Article
In: ACM Transactions on Intelligent Systems and Technology, vol. 2, no. 4, pp. 35:1–35:28, 2011.
@article{eps272015,
title = {Agent-based homeostatic control for green energy in the smart grid},
author = {Sarvapali Ramchurn and Perukrishnen Vytelingum and Alex Rogers and Nick Jennings},
url = {http://eprints.soton.ac.uk/272015/},
year = {2011},
date = {2011-01-01},
journal = {ACM Transactions on Intelligent Systems and Technology},
volume = {2},
number = {4},
pages = {35:1--35:28},
abstract = {With dwindling non-renewable energy reserves and the adverse effects of climate change, the development of the smart electricity grid is seen as key to solving global energy security issues and to reducing carbon emissions. In this respect, there is a growing need to integrate renewable (or green) energy sources in the grid. However, the intermittency of these energy sources requires that demand must also be made more responsive to changes in supply, and a number of smart grid technologies are being developed, such as high-capacity batteries and smart meters for the home, to enable consumers to be more responsive to conditions on the grid in real-time. Traditional solutions based on these technologies, however, tend to ignore the fact that individual consumers will behave in such a way that best satisfies their own preferences to use or store energy (as opposed to that of the supplier or the grid operator). Hence, in practice, it is unclear how these solutions will cope with large numbers of consumers using their devices in this way. Against this background, in this paper, we develop novel control mechanisms based on the use of autonomous agents to better incorporate consumer preferences in managing demand. These agents, residing on consumers' smart meters, can both communicate with the grid and optimise their owner's energy consumption to satisfy their preferences. More specifically, we provide a novel control mechanism that models and controls a system comprising of a green energy supplier operating within the grid and a number of individual homes (each possibly owning a storage device). This control mechanism is based on the concept of homeostasis whereby control signals are sent to individual components of a system, based on their continuous feedback, in order to change their state so that the system may reach a stable equilibrium. Thus, we define a new carbon-based pricing mechanism for this green energy supplier that takes advantage of carbon-intensity signals available on the internet in order to provide real-time pricing. The pricing scheme is designed in such a way that it can be readily implemented using existing communication technologies and is easily understandable by consumers. Building upon this, we develop new control signals that the supplier can use to incentivise agents to shift demand (using their storage device) to times when green energy is available. Moreover, we show how these signals can be adapted according to changes in supply and to various degrees of penetration of storage in the system. We empirically evaluate our system and show that, when all homes are equipped with storage devices, the supplier can significantly reduce its reliance on other carbon-emitting power sources to cater for its own shortfalls. By so doing, the supplier reduces the carbon emission of the system by up to 25% while the consumer reduces its costs by up to 14.5%. Finally, we demonstrate that our homeostatic control mechanism is not sensitive to small prediction errors and the supplier is incentivised to accurately predict its green production to minimise costs.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
With dwindling non-renewable energy reserves and the adverse effects of climate change, the development of the smart electricity grid is seen as key to solving global energy security issues and to reducing carbon emissions. In this respect, there is a growing need to integrate renewable (or green) energy sources in the grid. However, the intermittency of these energy sources requires that demand must also be made more responsive to changes in supply, and a number of smart grid technologies are being developed, such as high-capacity batteries and smart meters for the home, to enable consumers to be more responsive to conditions on the grid in real-time. Traditional solutions based on these technologies, however, tend to ignore the fact that individual consumers will behave in such a way that best satisfies their own preferences to use or store energy (as opposed to that of the supplier or the grid operator). Hence, in practice, it is unclear how these solutions will cope with large numbers of consumers using their devices in this way. Against this background, in this paper, we develop novel control mechanisms based on the use of autonomous agents to better incorporate consumer preferences in managing demand. These agents, residing on consumers' smart meters, can both communicate with the grid and optimise their owner's energy consumption to satisfy their preferences. More specifically, we provide a novel control mechanism that models and controls a system comprising of a green energy supplier operating within the grid and a number of individual homes (each possibly owning a storage device). This control mechanism is based on the concept of homeostasis whereby control signals are sent to individual components of a system, based on their continuous feedback, in order to change their state so that the system may reach a stable equilibrium. Thus, we define a new carbon-based pricing mechanism for this green energy supplier that takes advantage of carbon-intensity signals available on the internet in order to provide real-time pricing. The pricing scheme is designed in such a way that it can be readily implemented using existing communication technologies and is easily understandable by consumers. Building upon this, we develop new control signals that the supplier can use to incentivise agents to shift demand (using their storage device) to times when green energy is available. Moreover, we show how these signals can be adapted according to changes in supply and to various degrees of penetration of storage in the system. We empirically evaluate our system and show that, when all homes are equipped with storage devices, the supplier can significantly reduce its reliance on other carbon-emitting power sources to cater for its own shortfalls. By so doing, the supplier reduces the carbon emission of the system by up to 25% while the consumer reduces its costs by up to 14.5%. Finally, we demonstrate that our homeostatic control mechanism is not sensitive to small prediction errors and the supplier is incentivised to accurately predict its green production to minimise costs.
