Arguing about Trust in Multiagent Systems

Trust in multiagent systems is used for seeking to minimize
the uncertainty in the interactions among the agents. In this
paper, we discuss how to use argumentation to reason about
trust. Using the methodology of meta-argumentation, first we
represent the source of the information from which the argument
is constructed in the abstract argumentation framework
capturing the fact that b is attacked because b is from a particular
source s. We show how a source of information can
be attacked if it is not evaluated as trustworthy. Second, we
provide a fine grained representation of the trust relationships
between the information sources in which trust concerns not
only the sources but also the single arguments and attack relations
the sources propose. Moreover, we represent the evidences
in support of the arguments which are put forward by
the information sources and the agents can express arguments
by referring to other agents’ arguments. Meta-argumentation
allows us not to extend Dung’s abstract argumentation framework
by introducing trust and to reuse those principles and
properties defined for Dung’s framework.