Item


Beyond trust. Using fuzzy contextual filters for reliability assessment in multi agent systems

Trust modeling is widely recognized as an aspect of essential importance in the construction of agents and multi agent systems (MAS). As a consequence, several trust formalisms have been proposed over the last years. All of them have a limitation: they can determine the trustworthiness or untrustworthiness of the information received from a given agent, but they don鈥檛 supply mechanisms for correcting this information, in the case of it being unaccurate, in order to extract some utility from it. In order to overcome this limitation, this thesis introduces the concept of reliability as a generalization of trust, and presents Fuzzy Contextual Filters (FCF) as reliability modeling methods loosely based on system identification and signal processing techniques. This thesis illustrates their applicability to two domains: the appraisal variance estimation problem in the Agent Reputation and Trust (ART) testbed and Bar Systems, a class of optimization algorithms for reactive MAS

El modelatge de la confian莽a 茅s 脿mpliament reconegut com un aspecte d鈥檌mport脿ncia fonamental en la construcci贸 dels agents i de sistemes d鈥檃gents m煤ltiples (MAS). Com a conseq眉猫ncia, s鈥檋an proposat en els 煤ltims anys diversos formalismes de confian莽a. Tots ells tenen una limitaci贸: que puguin determinar la fiabilitat o la manca de credibilitat de la informaci贸 rebuda d鈥檜n agent determinat, per貌 no subministrar脿 els mecanismes per a la correcci贸 d鈥檃questa informaci贸, en el cas que sigui inexacte, per tal d鈥檈xtreure alguna utilitat ella. Per superar aquesta limitaci贸, aquesta tesi presenta el concepte de fiabilitat com una generalitzaci贸 de la confian莽a, i presenta Filtres contextuals Fuzzy (FCF) com a m猫todes de modelatge fiabilitat vagament basada en la identificaci贸 de sistemes i t猫cniques de processament de senyal. En aquesta tesi s鈥檌l路lustra la seva aplicaci贸 a dos dominis: el problema d鈥檈stimaci贸 de vari脿ncia d鈥檃valuaci贸 de la reputaci贸 d鈥檃gent i Trust (ART) banc de proves i Bar Systems, una classe d鈥檃lgorismes d鈥檕ptimitzaci贸 per reactiva MAS

Universitat de Girona

Manager: Rosa, Josep Llu铆s de la
Other contributions: Universitat de Girona. Departament d鈥橧nform脿tica i Matem脿tica Aplicada
Author: Acebo Pe帽a, Esteve del
Abstract: Trust modeling is widely recognized as an aspect of essential importance in the construction of agents and multi agent systems (MAS). As a consequence, several trust formalisms have been proposed over the last years. All of them have a limitation: they can determine the trustworthiness or untrustworthiness of the information received from a given agent, but they don鈥檛 supply mechanisms for correcting this information, in the case of it being unaccurate, in order to extract some utility from it. In order to overcome this limitation, this thesis introduces the concept of reliability as a generalization of trust, and presents Fuzzy Contextual Filters (FCF) as reliability modeling methods loosely based on system identification and signal processing techniques. This thesis illustrates their applicability to two domains: the appraisal variance estimation problem in the Agent Reputation and Trust (ART) testbed and Bar Systems, a class of optimization algorithms for reactive MAS
El modelatge de la confian莽a 茅s 脿mpliament reconegut com un aspecte d鈥檌mport脿ncia fonamental en la construcci贸 dels agents i de sistemes d鈥檃gents m煤ltiples (MAS). Com a conseq眉猫ncia, s鈥檋an proposat en els 煤ltims anys diversos formalismes de confian莽a. Tots ells tenen una limitaci贸: que puguin determinar la fiabilitat o la manca de credibilitat de la informaci贸 rebuda d鈥檜n agent determinat, per貌 no subministrar脿 els mecanismes per a la correcci贸 d鈥檃questa informaci贸, en el cas que sigui inexacte, per tal d鈥檈xtreure alguna utilitat ella. Per superar aquesta limitaci贸, aquesta tesi presenta el concepte de fiabilitat com una generalitzaci贸 de la confian莽a, i presenta Filtres contextuals Fuzzy (FCF) com a m猫todes de modelatge fiabilitat vagament basada en la identificaci贸 de sistemes i t猫cniques de processament de senyal. En aquesta tesi s鈥檌l路lustra la seva aplicaci贸 a dos dominis: el problema d鈥檈stimaci贸 de vari脿ncia d鈥檃valuaci贸 de la reputaci贸 d鈥檃gent i Trust (ART) banc de proves i Bar Systems, una classe d鈥檃lgorismes d鈥檕ptimitzaci贸 per reactiva MAS
Document access: http://hdl.handle.net/2072/295083
Language: eng
Publisher: Universitat de Girona
Rights URI: http://creativecommons.org/licenses/by/3.0/es/
Subject: Tesis i dissertacions acad猫miques
Fuzzy contextual filters
Filtres difusos contextuals
Filtros difusos conceptuales
Crew intelligence
Multiagent systems
Sistemes multiagent
Sistemas multiagente
Bar systems
004 - Inform脿tica
Title: Beyond trust. Using fuzzy contextual filters for reliability assessment in multi agent systems
Type: info:eu-repo/semantics/doctoralThesis
Repository: Recercat

Subjects

Authors