Findings and Discussions on the Neural Trust and Multi-Agent System

Findings and Discussions on the Neural Trust and Multi-Agent System

Gehao Lu (University of Huddersfield, UK & Yunnan University, China) and Joan Lu (University of Huddersfield, UK)
Copyright: © 2017 |Pages: 23
DOI: 10.4018/978-1-5225-1884-6.ch020

Abstract

This chapter focuses on the testing for a complete systematic neural trust model developed previously based on the trust learning algorithms, trust estimation algorithm and reputation mechanisms. The focus is to describe the detailed design of the model and explain the rationales behind the model design. The purpose is to evaluate the proposed neural trust model from different aspects and analyze the results of the evaluations. Experiments have been conducted. Results are presented and discussed. Finally, based on the analysis and comparison of acquired results, conclusions are drawn.
Chapter Preview
Top

2. Test-Bed Design And Implementation

For evaluating the effectiveness of the proposed computational trust and reputation models, a test bed is designed and implemented to provide an experimental platform. The test bed is built in Java programming language and it makes use of the agent framework JADE (Bellifemine, Caire & Greenwood, 2007) to provide agent management, agent communication and multi-agent environment. The test bed is based on the JADE API and JADE Agent Containers. The research chooses the experimental scenario as the electronic commerce activities that are finished by autonomous agents.

Complete Chapter List

Search this Book:
Reset