Elisa Burato
Meaning Negotiation in Multiple Agent Systms: An Automated Reasoning Approach

Meaning Negotiation is the process by which two or more than two agents discuss about the meaning of a set of terms in order to find an agreement about it. Meaning Negotiation begins with inconsistent viewpoints of the agents. This mechanism of reaching an agreement is important in automating intelligent sy- stems that behave on behalf of humans (e-Commerce, e-Learning, Finders, etc.). Meaning Negotiation was widely studied by the Artificial Intelligence community but, to the best of my knowledge, there are no formalisms modelling the Meaning Negotiation that are independently of the number of participants and of their expression language. Moreover, the current approaches deal with Meaning Negotiation as a process in which the issue is “quantitative”, i.e. it can be divided, increased or decreased numerically. The issue of Meaning Negotiation is knowledge, thus it is straightforward to model its negotiation since it is not immediately evident how to divide, to increase and to decrease knowledge. In this thesis, I make a survey of the current Meaning Negotiation approaches and formalise it in a game theoretic perspective to model the interaction protocol, and in a deductive manner to represent the reasoning mechanism of agents. Moreover, I assume and formalise the strategical behaviour of negotiation’s participants in Defeasible Logic. This pathway brings to a comprehensive formalism, that, within he game theoretic endeavour, describes the process, the behaviour of the agents, their goals, and the governing mechanism of Meaning Negotiation in a compact fashion.

(pdf, BiBTeX)