GAVEL: a sanction-based regulation mechanism for normative multiagent systems

Detalhes bibliográficos
Ano de defesa: 2019
Autor(a) principal: Lima, Igor Conrado Alves de
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: eng
Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: https://www.teses.usp.br/teses/disponiveis/45/45134/tde-06072020-170444/
Resumo: The use of a normative approach to govern multiagent systems (MAS) has been motivated by the increasing interest in balancing between agents\' autonomy and global system control. In normative multiagent systems (NMAS), despite the existence of norms specifying rules about how agents ought or ought not to behave, agents have the autonomy to decide whether or not to act in compliance with such norms. A suitable way to govern agents is using sanction-based enforcement mechanisms. These mechanisms allow agents autonomy while maintaining a certain system control level through sanction applications. However, most enforcement mechanisms found in the literature lack support for the association of norm compliance or violation to different sanction categories and strength, thus refraining agents from sanction reasoning and decision capabilities. An exception is the model proposed by Nardin et al. (2016). Based on this latter model, this thesis presents an operational sanctioning enforcement framework, named GAVEL, which endows agents with the capability to decide for the most appropriate sanctions to apply, depending on their context assessed by a set of decision factors. The advantages of applying different sanctions categories and strength are illustrated by a case study from the domain of economics, namely the public goods game (PGG). The experimental results show that allowing agents to decide between material and social sanctions leads to similar cooperation rates but greater wealth levels in comparison to solely using material sanction.