Multi-agent systems applications in a number of areas such as e-commerce, disaster management, and information acquisition through embedded devices (e.g. wireless sensor networks) have generated a number of new challenges for algorithm designers. These challenges mainly take the form of very hard optimisation problems that are substantially different from problems traditionally dealt with in other areas (e.g. industrial processes or scheduling applications). More specifically, novel challenges come from the distributed nature of multiagent systems where the actors reside on different computational units and can communicate only a limited amount of information with their neighbours. Moreover, the agents may be acting on behalf of different stakeholders each with its own aims and objectives, have different computation/communication capabilities, and be tied to physical devices prone to failures. Moreover, given the dynamic nature of the application scenarios, effective algorithms have to provide anytime solutions and approximate techniques are often required/desirable.

We focus on the design of techniques for coordinating large populations of agents by modelling the coordination problem as a distributed constraint optimisation problem (DCOP). In particular, we mainly focus on:

  • the design of complete algorithms
  • the design of approximate algorithms that provide anytime solutions
  • assessing quality guarantees both prior to the execution of an algorithm (at design time) and during the execution of an algorithm (run time)

Collectiveware: Highly-parallel algorithms for collective intelligence

Initial/final date: 
16 June 2017 to 15 June 2019
Project researchers: 
Main researcher: 
Project type: 
International
Funding Entity: 
European Commission
Description: 

In recent years, more and more scenarios pose challenges that require collective intelligence solutions based on networks (knowledge networks, social networks, sensor networks). New forms of collaborative consumption, collaborative making, collaborative production, all rely on a common task, the formation of collectives. This task is crucial in many real-world applications domains. Notable examples of actual-world collective formation scenarios are Collective Energy Purchasing (CEP), a collaborative consumption scenario, and Team Formation (TF), a collaborative production scenario. Within the Artificial Intelligence literature, current state of the art algorithms cannot provide the level of scalability and the solution quality required by actual-world collective formation problems, hence novel algorithms are needed to tackle these problems. To achieve this objective, we aim at proposing novel algorithms that are capable to exploit modern highly-parallel architectures. On the one hand, highly-parallel architectures have been successfully applied in many different scenarios so to achieve tremendous performance improvements. These advancements encourage the investigation of parallelisation also in collective formation, with the objective of achieving the same benefits. On the other hand, our past research indicates that considering the structure of the collective formation problem leads to notable benefits in terms of scalability and solution quality. Thus, we propose to take a novel algorithmic design approach that considers both the structure of the scenario and at the same time exploits modern highly-parallel architectures. Our algorithms will be evaluated in two prominent collective intelligence application domains: the CEP and TF domains. The choice of these two application domains will serve to show the generality of our algorithmic design approach, since they are representative of two structurally different families of actualworld collective formation problems.

Project Status: 
Ongoing
Funding Amount (€): 
158121.00
Research line: 
Optimisation
Acronym: 
HPA4CF
Geographical scope: 
European
Grant type: 
Competitive
Transferència: 

Tecnologías para potenciar colectivos humanos en la red eléctrica inteligente

Initial/final date: 
01 January 2016 to 31 December 2018
Project researchers: 
Main researcher: 
Project type: 
Plan Nacional
Funding Entity: 
Ministerio de Economía y Competitividad
Description: 

Nowadays, information and communication technologies empower individuals’ communication, but are far from exploiting all their collavorative potential. Despite the recent surge of interest in technologies to help individuals collaborate, such as crowdsourcing, they are still in its infancy. And yet, such technologies are core to enable future smart infrastructures for the citizens. Indeed, participation and social innovation are at the heart of the development of emerging infrastructures such as the smart city or the smart grid, which require intensive collaboration within collectives. Here we focus on microgids, localized groupings of electricity resources and loads, which can function autonomously. We target novel technologies that empower human collectives to operate micro-grids to achieve sustainable energy management by supporting their self-awareness, cooperation, and self-governance.

We argue that collaboration in micro-grids can be articulated either through market-based energy allocation mechanisms aimed at supporting local trading, or by means of norm-based mechanisms agreed upon by a micro-grid community’s prosumers (those that consume and produce energy).

On the one hand, we will introduce a novel peer-to-peer market to allow prosumers to individually trade electricity over a micro-grid while satisfying grid transmission constraints. This market will favour the decentralized generation of renewable energy and the reduction of inefficiencies in current centralized model of production and transmission (in terms of CO2 reduction). Energy allocation will be based on fixed market rules aimed at cost reduction and it relies on cooperation of prosumers to relay energy from one to another.

On the other hand, we will introduce a novel social computing approach to allow a community’s prosumers to agree on the rules to employ to allocate energy. Unlike our market-based approach, this social approch will be driven by consensus, hence fostering community engagement, social cohesion, and a more democratic energy system. Our purpose is that prosumers decide themselves what they consider as fair energy allocation rules. Rephrasing Nobel-prize winner E. Ostrom, involving a community’s participants in their decisions improves its long term operation, particularly when resources are limited. Thus, Ostrom’s principles are expected to ease energy sustainability along time. Furthermore, we will exploit location-based social networks to assess whether a community values the rules of current allocation system or if some changes are required.

Complementarily, we will also apply gamification techniques to educate community prosumers so that they learn efficient energy management practices. Our assumption is that responsible prosumers are expected to better manage energy, and thus help achieve larger CO2 reductions.

Upon the proposal of our allocation approaches, we will analyse them empirically with both synthetic data and with humans in controlled lab scenarios. We will also quantify the potential benefits of gamification to yield more efficient user energy profiles.

This project tackles the “Economía y Sociedad digital” and “Energía segura, eficiente y limpia” challenges. Overall, we expect our project to have socio-economic impact by providing new ways to organize the energy system driven by a sense of community and local ownership. Sustainability will be rooted in education, cooperation and social innovation.

Referencia: 

Ref. TIN2015-66863-C2-1-R (MINECO/FEDER)

Project Status: 
Ongoing
Funding Amount (€): 
63646.00
Research line: 
Optimisation
Acronym: 
Collectiveware
External researchers: 
Sarvapali Ramchurn
Michael Wooldridge
Mark Klein
Gauthier Picard
Maite López-Sánchez
Inmaculada Rodríguez
Geographical scope: 
National
Grant type: 
Competitive
Transferència: 

AUTONOMIC ELECTRONIC INSTITUTIONS

Initial/final date: 
01 October 2006 to 30 September 2009
Main researcher: 
Project type: 
Plan Nacional
Funding Entity: 
TIN2006-15662-C02-01
Description: 
Electronic institutions (EIs) allow to establish interaction conventions among agents --persons and/or programs--on the basis of a distributed, open and computationally dynamic environment. This project has as a main goal the study of techniques which allow to provide EIs with autonomic capabilities that allow them to offer a dynamic response under changing circumstances by adopting interaction conventions, producing a high-level development environment for autonomous electronic institutions (AEI). From all those features characterizing an autonomous system, we will focus mainly on the study of auto configuration and reconfiguration. Besides, the development environment will provide us with tools for specification and analysis of AEI which will allow us to reduce the time and complexity of their development. In order to validate the resulting framework, we will carry out experiments based on real-world problems modelled by means of AEIs. More information: http://e-institutions.iiia.csic.es
Funding Amount (€): 
220000.00
Phd Students: 
Norman Salazar-Ramirez
Meritxell Vinyals
Tomas Trescak
Mariela Morveli Espinoza
José Luis Fernández
Research line: 
Optimisation
Acronym: 
IEA
External researchers: 
Jesús Cerquides,Maite López-Sánchez,