Multi-User Cooperative Computation Framework Based on Bertrand Game

TL;DR

This paper proposes a cooperative computing framework for mobile users utilizing neighboring resources, employing a Bertrand game-based incentive scheme to optimize resource sharing and payments, with proven effectiveness through simulations.

Contribution

It introduces a novel Bertrand game-based incentive scheme for multi-user cooperative computing, including closed-form NE solutions and a distributed algorithm for incomplete information scenarios.

Findings

Closed-form Nash equilibrium under complete information

Distributed iterative algorithm for incomplete information

Simulation results verify scheme effectiveness

Abstract

In this paper, a multi-user cooperative computing framework is applied to enable mobile users to utilize available computing resources from other neighboring users via direct communication links. An incentive scheme based on Bertrand game is proposed for the user to determine \textit{who} and \textit{how} to cooperate. We model the resource demand users as \textit{buyers} who aim to use minimal payments to maximize energy savings, whereas resource supply users as \textit{sellers} who aim to earn payments for their computing resource provision. A Bertrand game against \textit{buyer's market} is formulated. When the users have \textit{complete information} of their opponents, the Nash equilibrium (NE) of the game is obtained in closed form, while in the case of \textit{incomplete information}, a distributed iterative algorithm is proposed to find the NE. The simulation results verify the effectiveness of the proposed scheme.