A fault-tolerant cooperative distributed system

Abstract

The paper proposes a highly distributed environment for Cooperating Knowledge Based Systems (CKBSs), in which autonomous knowledge based systems, called agents, cooperate/compete together to solve problems of real-world distributed applications. The paper then goes on to explore how such an environment can be exploited to provide fault tolerance, particularly fault localisation, and the replacement of faulty components followed by rescheduling, providing a resilient and stable operational environment.