Matching Markets for Spectrum Sharing

10 Pages Posted: 31 Mar 2017 Last revised: 16 Aug 2017

See all articles by Marcela Gomez

Marcela Gomez

University of Pittsburgh - School of Information Sciences

Martin B. H. Weiss

University of Pittsburgh - School of Computing and Information

Giulia McHenry

Government of the United States of America - National Telecommunications and Information Administration (NTIA)

Linda Doyle

Trinity College (Dublin)

Date Written: March 31, 2017


Next generation networks aim at improving connectivity and capacity, adding to the current range of available services and expanding their reachability. For these systems to work, they need to be compatible with legacy technologies in addition to making use of (limited) available spectrum resources. This is one of the reasons why spectrum sharing has been at the forefront of the list of enablers for such systems. From federal-commercial sharing to finding opportunities in millimeter-wave spectrum, we have witnessed the formulation of multiple approaches to making spectrum sharing happen.

Existing work on spectrum sharing is wide ranging and includes technological as well as market based approaches. Spectrum markets settings have called for different definitions of spectrum-related resources as a means to increase market thickness and thus improve the opportunities for finding market liquidity. In a similar way, we find proposals of network models which aim at adapting technical definitions of spectrum resources, such as those that are the product of virtualization. In this work, we adopt a market perspective for spectrum sharing within the context of more comprehensive network definitions such as those envisioned for next generation networks.

The particular network model we build upon is that introduced by Doyle et al. in [1]. This framework envisions heterogeneous physical networks that collaborate through virtualization to provide a consistent service to end users. Such an approach suggests three main participating entities: Resource Providers, Virtual Network Builders and Service Providers. The Resource Providers (RPs) are current resource owners or incumbents who can make their excess resources available in a common pool. The Service Providers (SPs) are new market entrants or existing providers who need additional resources in order to meet the demand of their customers (i.e., end users). The Virtual Network Builders (VNBs) are intermediaries (or “middlemen") whose function is to obtain resources from the pool according to the demand of a subset of SPs whom they consider their customers. These entities and their definitions aim at outlining what a next-generation network would look like and what it would require.

Our objective in this work is to build a model representing a market setting that would be compatible with this kind of network. The theoretical framework for this model is both “middleman" theory [2] and “matching markets" [3, 4, 5]. We model VNB - SP interactions based on real-world middlemen or brokers; in fact, we envision a partnership-forming process to take place between these two sets of entities. The modeled matching process takes into account parameters that are relevant to SPs and VNBs in order to define their preferences and thus form acceptable matches. This method further allows us to observe how the preference parameters can influence the matching outcome. In other words, by applying different weights to these criteria we can observe whether there are changes on the resulting number of partnerships formed; the percentage of geographical demand that becomes market demand; the range of lump sum fees requested by and paid to VNBs, among other factors.

We employ an agent-based model written in MATLAB to explore the function and performance of this model. MATLAB offers the necessary tools for handling the data we expect our agents to utilize and the results generated by the model. For result analysis, we rely on Python and R, due to the additional tools provided by these platforms for data processing and analysis. This modeling approach allows us to examine the benefits and constraints that novel network configurations entail. Particularly, our results can be useful to determine the drivers for resource sharing under the proposed configuration. In addition, we can formulate recommendations that could be extrapolated to other proposed sharing schemes, which include similar network participants and settings.

References  [1] L. Doyle, J. Kibilda, T. K. Forde, and L. DaSilva, \Spectrum Without Bounds, Networks Without Borders," Proceedings of the IEEE, vol. 102, no. 3, pp. 351-365, mar 2014. [2] M. Krakovsky, The Middleman Economy. Palgrave Macmillan, 2015. [3] A. E. Roth and M. A. O. Sotomayor, Two-sided matching: A study in game-theoretic modeling and analysis. Cambridge University Press, 1992, no. 18. [4] A. E. Roth, \What Have We Learned from Market Design?" The Economic Journal, vol. 118, no. 527, pp. 285-310, 2008. [5] A. E. Roth, Who Gets What and Why. Houghton Mifflin Harcourt Publishing Company, 2015.

Keywords: Matching Markets, Next-generation networks, Spectrum Sharing, Agent-based Modeling

Suggested Citation

Gomez, Marcela and Weiss, Martin B. H. and McHenry, Giulia and Doyle, Linda, Matching Markets for Spectrum Sharing (March 31, 2017). Available at SSRN:

Marcela Gomez (Contact Author)

University of Pittsburgh - School of Information Sciences ( email )

United States

Martin B. H. Weiss

University of Pittsburgh - School of Computing and Information ( email )

135 N Bellefield Ave
Pittsburgh, PA 15260
United States

Giulia McHenry

Government of the United States of America - National Telecommunications and Information Administration (NTIA) ( email )

1401 Constitution Avenue, N.W.
Washington, DC 20230
United States

Linda Doyle

Trinity College (Dublin) ( email )

Dublin, D2

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics