Pricing and Equilibrium Analysis of Network Market Systems
Date
2018-01-18Metadata
Show full item recordAbstract
Markets have been the most successful method of identifying value of goods and services.
Both large and small scale markets have gradually been moving into the Internet domain,
with increasingly large numbers of diverse participants. In this dissertation, we consider several
problems pertaining to equilibria in networked marketplaces under different application
scenarios and market sizes. We approach the question of pricing and market design from two
perspectives. On the one hand, we desire to understand how self-interested market participants would set prices and respond to prices resulting in certain allocations. On the other
hand, we wish to evaluate how best to allocate resources so as to attain efficient equilibria.
There might be a gap between these viewpoints, and characterizing this gap is desirable.
Our technical approaches follow the number of market participants, and the nature of
trades happening in the market. In our first problem, we consider a market of providing
communication services at the level of providing Internet transit. Here, the transit Internet
Service Provider (ISP) must determine billing volumes and set prices for its customers who
are _rms that are content providers, sinks, or subsidiary ISPs. Demand from these customers
is variable, and they have different impacts on the resources that the transit ISP needs to
provision. Using measured data from several networks, we design a fair and flexible billing
scheme that correctly identifies the impact of each customer on the amount of provisioning
needed.
While the customer set in the first problem is finite, many marketplaces deal with a very
large number of agents that each have ephemeral lifetimes. Here, agents arrive, participate in
the market for some time, and then vanish. We consider two such markets in such a regime.
The first is one of apps on mobile devices that compete against each other for cellular data
service, while the second is on service marketplaces wherein many providers compete with
each other for jobs that consider both prices and provider reputations while making choices
between them. Our goal is to show that a Mean Field Game can be used to accurately
approximate these systems, determine how prices are set, and characterize the nature of
equilibria in such markets.
Finally, we consider efficiency metrics in large scale resource sharing networks in which
bilateral exchange of resources is the norm. In particular, we consider peer-to-peer (P2P)
file sharing under which peers obtain chunks of a file from each other. Here, contrary to
the intuition that chunks must be shared whenever one peer has one of value to another, we
show that a measure of suppression is needed to utilize resources efficiently. In particular, we
propose a simple and stable algorithm entitled Mode suppression that attains near optimal
file sharing times by disallowing the sharing of the most frequent chunks in the system.
Citation
Reddyvari Raja, Vamseedhar Reddy (2018). Pricing and Equilibrium Analysis of Network Market Systems. Doctoral dissertation, Texas A & M University. Available electronically from https : / /hdl .handle .net /1969 .1 /173298.