Generalized multi-cluster game under partial-decision information with applications to management of energy internet

Abstract

The decision making and management of many engineering networks involves multiple parties with conflicting interests, while each party is constituted by multiple agents. Such problems can be cast as a multi-cluster game. Each cluster is treated as a self-interested player in a non-cooperative game where agents in the same cluster cooperate to optimize the cost function of the cluster. In a large-scale network, the information of agents in a cluster can not be available immediately to agents beyond this cluster, which raises challenges to the existing Nash equilibrium seeking algorithms. Hence, we consider a partial-decision information scenario in multi-cluster games. We reformulate the problem by finding zeros of the sum of preconditioned monotone operators by the primal-dual analysis and graph Laplacian matrix. Then a distributed generalized Nash equilibrium seeking algorithm is proposed without requiring full awareness of its opponent clusters’ decisions based on a forward-backward-forward method. With the algorithm, each agent estimates the strategies of all the other clusters by communicating with neighbors via an undirected network. We show that the derived operators can be monotone when the communication strength parameter is sufficiently large. We prove the algorithm convergence by providing a sufficient condition of the fixed point theory. We discuss its potential application in Energy Internet with numerical studies.