Merging Existential Rules Programs in Multi-Agent Contexts through Credibility Accrual

CRISTHIAN A. D. DEAGUSTINI; JUAN CARLOS TEZE; M. VANINA MARTÍNEZ; MARCELO A. FALAPPA; GUILLERMO R. SIMARI

INFORMATION SCIENCES

ELSEVIER SCIENCE INC

Lugar: Amsterdam; Año: 2021 vol. 555 p. 236 - 259

0020-0255

Merging operators represent a signicant tool to extract a consistent and informative view from a set of agents. The consideration of practical scenarios where some agents can be more credible than others has contributed to substantially increase the interest in developing systems working with trust models. In this context, we propose an approach to the problem of merging knowledge in a multiagent scenario where every agent assigns to other agents a value reflecting its perception on how credible each agent is. The focus of this paper is the introduction of an operator for merging Datalog ontologies considering agents´credibility. We present a procedure to enhance a conflict resolution strategy by exploiting the credibility attached to a set of formulas; the approach is based on accrual functions that calculate the value of formulas according to the credibility of the agents that inform them. We show how our new operator can obtain the best-valued knowledge base among consistent bases available, according to the credibilities attached to the sources.