D-S Theory-based Trust Model FIRE^+ in Multi-agent Systems

Abstract

In open multi-agent systems (MAS), interaction between agents is uncertain. Trust plays an important role in deciding with whom to interact and how to interact with it. Current trust models have solved the problem of how to quantify, deduce and evaluate trust, but there are still challenges. This paper presents FIRE <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">+</sup> trust model using utility loss of issue to evaluate trust values. With the representation of matrix, it efficiently increases the rating information. Also by introducing D-S theory and WMA, it can handle uncertain information and dishonest witness finely. Meanwhile, our approach introduces the concept of information-amount to derive each issue 's weight so as to gain the whole trust value of the target agent. It also takes witness reliability into account. Experiments show that FIRE <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">+</sup> is more effective than others.