MINE GOLD to Deliver Green Cognitive Communications

TL;DR

This paper introduces MINE GOLD, a model for spectrum information markets in TV white space networks, analyzing how multiple databases and devices interact to optimize profits and system efficiency in a green cognitive communication context.

Contribution

It develops a hierarchical game-theoretic model for multi-database spectrum information markets, analyzing pricing strategies and market equilibrium in TV white space networks.

Findings

Database revenues are non-monotonic with the number of databases.

Positive network externalities enhance revenues and system performance.

Equilibrium behaviors depend on system parameters and pricing strategies.

Abstract

Geo-location database-assisted TV white space network reduces the need of energy-intensive processes (such as spectrum sensing), hence can achieve green cognitive communication effectively. The success of such a network relies on a proper business model that provides incentives for all parties involved. In this paper, we propose MINE GOLD (a Model of INformation markEt for GeO-Location Database), which enables databases to sell the spectrum information to unlicensed white space devices (WSDs) for profit. Specifically, we focus on an oligopoly information market with multiple databases, and study the interactions among databases and WSDs using a two-stage hierarchical model. In Stage I, databases compete to sell information to WSDs by optimizing their information prices. In Stage II, each WSD decides whether and from which database to purchase the information, to maximize his benefit of using the TV white space. We first characterize how the WSDs' purchasing behaviors dynamically evolve, and what is the equilibrium point under fixed information prices from the databases. We then analyze how the system parameters and the databases' pricing decisions affect the market equilibrium, and what is the equilibrium of the database price competition. Our numerical results show that, perhaps counter-intuitively, the databases' aggregate revenue is not monotonic with the number of databases. Moreover, numerical results show that a large degree of positive network externality would improve the databases' revenues and the system performance.