Multi-source multiple change on belief bases

TAMARGO, LUCIANO H.; DEAGUSTINI, CRISTHIAN A.D.; GARCÍA, ALEJANDRO J.; FALAPPA, MARCELO A.; SIMARI, GUILLERMO R.

INTERNATIONAL JOURNAL OF APPROXIMATE REASONING

ELSEVIER SCIENCE INC

Año: 2019 vol. 110 p. 145 - 163

0888-613X

In a multi-agent system, an agent?s belief base can be modified when that agent receives information from other agents. Here, we study the case when these belief base changes occur in a collaborative multi-agent system environment, and the agents in the system exhibit different degrees of credibility; the former assumption guarantees that the agents are truthful and they help each other, and the latter introduces the opportunity of choosing the best possible source of information.Belief Revision in Multi-Agent Systems has been analyzed in [1] where the distinction between Multi-Agent Belief Revision(MABR) and Individual Belief Revision(IBR) is explained. On the one hand, MABR[1?4]studies the overall belief revision be-havior of agent teams in which, to pursue the mutual goal, the involved agents need to communicate, cooperate, coordinate, and negotiate with one another. On the other hand, IBR is divided into two different areas: belief revision in a single agent environment called Single agent Belief Revision(SBR)[5] and individual belief revision in a multi-agent environment called Multi-Source Belief Revision(MSBR). In this last area, an individual belief revision process is carried out in a multi-agent envi-ronment where the new information may come from multiple sources that may be in conflict. We can also distinguish two types of MSBR: unitary MSBRwhere an agent can receive from different informant agents an atomic piece of information (see for instance[6?8]), and conjunctive MSBRwhere the received information is a set of objects, probably provided by one or more sources. Our proposal is focused on the latter approach.