Ryan Thomas’ dissertation talks about the advantages of implementing a cognitive network. Unlike many of the networks and game theory methods discussed in class this network runs on the notion of what is best for all parties involve instead or being centralized on each individual node in the network. The concept of learn and rationalizing about what each decisions impact would have on the network as a whole strips away Nash Equibrilums (NE) and brings to light the optimal situation for the system. Ryan Thomas splits cognitive networks that require selfish strategies into two classes saying,

“The first class we identify is the potential class, which assures the convergence of the network to NE that are local-optima for network objectives. The second class we identify is the quasi-concave class, which assures the convergence of the network to a Pareto Optimal Nash Equilibrium (PONE) that is both a network and cognitive element optima.” [p. 59]

He goes on to state that when constructing a cognitive network one must factor the three critical design decisions (price of selfishness, ignorance and control) that will ultimately affect the performance.

The layout provided by Thomas and his colleagues gives a foundation for combating the problems brought about by the increase in file size and bandwidth-intensive applications that many users have become accustom to. With the rise of file sharing and online software, it has become important to keep people better connected with their data. Cognitive networks will give intelligence to networks, causing the most basic tasks of packet routing to the most complex the ability to utilize the network while not crippling another process from essential resources.

Link:

Ryan Thomas’ Dissertation