Reem Bahgat

This work presents a model for assigning trust values to agents operating within a collaborative multi-agent system. The model enables agents to assess the trustworthiness of their peers, and thus, to be able to select reliable ones for cooperation and coalition formation. In this work, the performance of a group of agents – a team – that collaborate to achieve a shared goal where the individual contribution of each agent is unknown, is evaluated. The work thus aims to present a reliable method for calculating a trust value for agents involved in teamwork. More specifically, this research presents a model – called Inducing the Trust Ratio Model - for evaluating the individual trustworthiness of a group of agents. Toward this end, the model makes use of genetic algorithms to induce the trust ratio of each coalition member. Empirical analysis is undertaken to evaluate the effectiveness of this model