Distributed Nash Equilibrium Seeking in N-Cluster Games with Fully Uncoordinated Constant Step-Sizes

TL;DR

This paper introduces a distributed Nash equilibrium seeking algorithm for N-cluster games that allows agents to independently choose their step-sizes, ensuring convergence under certain conditions, and broadening the applicability of NE seeking methods.

Contribution

It proposes a novel NE seeking algorithm for N-cluster games with fully uncoordinated constant step-sizes, enabling more flexible and practical agent updates.

Findings

Agents' decisions converge linearly to NE with small step-size heterogeneity.

The algorithm works with fully uncoordinated step-sizes across agents.

Numerical example confirms theoretical convergence results.

Abstract

Distributed optimization and Nash equilibrium (NE) seeking problems have drawn much attention in the control community recently. This paper studies a class of non-cooperative games, known as N-cluster game, which subsumes both cooperative and non-cooperative nature among multiple agents in the two problems: solving distributed optimization problem within the cluster, while playing a non-cooperative game across the clusters. Moreover, we consider a partial-decision information game setup, i.e., the agents do not have direct access to other agents' decisions, and hence need to communicate with each other through a directed graph. To solve the N-cluster game problem, we propose a distributed NE seeking algorithm by a synthesis of consensus and gradient tracking. Unlike other existing discrete-time methods for N-cluster games where either a common step-size is publicly known by all agents or only known by agents from the same cluster, the proposed algorithm can work with fully uncoordinated constant step-sizes, which allows the agents (both within and across the clusters) to choose their own preferred step-sizes. We prove that all agents' decisions converge linearly to their corresponding NE so long as the largest step-size and the heterogeneity of the step-sizes are small. We verify the derived results through a numerical example in a Cournot competition game.