Cyber attack-resilient consensus algorithm for cooperative multi-agent system

Abstract

This paper proposes a task assignment algorithm in multi-agent system and suggestion to secure task assignment algorithm against malicious cyber-attack. The goal of the task assignment algorithm in a multi-agent system is to maximize optimality of the solution. Previous studies of multi agent coordination have focused on how to assign tasks, and control the system. Consensus Based Bundle Algorithm(CBBA) is a decentralized plan consensus based auction algorithm guaranteeing feasible, conflict-free solutions. In this thesis, the issue of secure task allocation for detecting attacks on plan auction and information communication is addressed. The aim of this study is to make task assignment process secure. Assignment simulations showed that the proposed algorithm can be used as a detection scheme against cyber-attack in decentralized multi-agent system.