DISTRIBUTED INTERVAL CONSENSUS IN NETWORKED SYSTEMS WITH PRIVACY PRESERVATION

Abstract Interval consensus is an important generalization of conventional consensus problems by allowing each agent to individually nominate an acceptable interval for their consensus value. However, as other consensus problems, agents in the network exchange information explicitly among neighbours and disclose their values without any protection for sensitive information causing serious privacy concerns in many applications in distributed multiagent systems. We propose a privacy-preserving approach consisting of decomposition and weighting mechanisms. Based on this approach, we show that the agents in the network can achieve interval consensus with the final consensus value within the intersection of all proposed intervals if the intersection is nonempty and the network is connected. Moreover, the privacy of the initial states of the agents is guaranteed against internal and external adversaries. The proposed consensus protocol is simple and efficient, and it can be implemented in a distributed manner over the network.