Trustworthy Distributed Average Consensus

This paper proposes a distributed algorithm for average consensus in a multi-agent system under a fixed, possibly directed communication topology, in the presence of malicious agents (nodes) that may try to influence the average consensus value by manipulating their initial values and/or their updates in an arbitrary manner. The proposed algorithm is iterative and asymptotically converges to the average of the initial values of the non-malicious nodes (referred to as the average of the trustworthy nodes), as long as the underlying topology that describes the information exchange among the non-malicious nodes is strongly connected. The algorithm assumes that each node receives (at each iteration or periodically) side information about the trustworthiness of the other nodes, and it uses such trust assessments to determine whether or not to incorporate messages received from an in-neighbor or take into account, for its updates and transmissions, a particular out-neighbor. The algorithm allows the perceived trustworthiness of a node about another node to be asymmetric and to fluctuate during the iteration, and guarantees asymptotic convergence to the average of the trustworthy nodes, as long as the trust assessments for each non-malicious node eventually reflect correctly the status (malicious or non-malicious) of its neighboring nodes. Keywords: Distributed averaging, multi-agent systems, fault-tolerant consensus, resilience, trustworthy computation, trust values.