Engineering autonomous trust-management requirements for software agents: Requirements and concepts

Abstract

Due to the need for open and distributed computer systems the demand for trust and autonomous trust management by software agents has become an important topic. This led to the development of different frameworks to support trust management by autonomous software agents. These frameworks enable software agents to autonomously form trust, make trust decisions, and update trust, but are often limited to certain application domains. Currently there has been little effort (e. g. [1], [2]) to integrate trust modeling into phases and models of Agent Oriented Software Engineering (AOSE). So one step to further fit AOSE methodologies to develop open agent societies is to integrate a systematic approach to develop trust models. This paper describes a systematic approach to identify requirements on software agents in order to autonomously manage trust. For this purpose the extendable agent-oriented development process of O-MaSE (Organization-based Multi-Agent System Engineering) and its modeling concepts are extended. © Springer Science+Business Media B.V. 2010.