New Media Research Studio

Required Reading Summary:

Dr Anna Nagurney
The Virtual Center for Supernetworks- University of Massachusetts
Networks The Science that Spans Disciplines

She breaks it up into the following topics so I guess I’ll do the same:
-Background
-Examples of Physical Networking
-Network Components
-Scientific Study of Networks
-Classical Networks and Applications
-Interdisciplinary Impact of Networks
-Characteristics of Networks Today
-The Braess Paradox
-Supernetworks
-Novel Applications- Financial Networks to Social Networks

BACKGROUND
Networks are a part of the way that our society and economy survive in terms of communication, production, and transportation. Networks, she explains, shouldn’t just be thought of in terms of technology, it’s a system that runs everything around us including. A network is a way of connecting humans and their ideas.

EXAMPLES OF PHYSICAL NETWORKS
It’s an ancient idea that runs all the way back to congestion traffic management of chariots. She also talks about energy networks, which are studied within the context of transportation and communication networks.

NETWORK COMPONENTS
The basic components of networks are:
nodes: transportation intersections
links or arcs: bi-directional, connections without direction, directional
just dealing with nodes and links is graphs, but networks also deal with flows.

Going back to the transportation example: a node is the intersections, airport, workplace
Link is the road, airline route, or railroad track
Flow is the automobile, trains, or planes

SCIENTIFIC STUDY OF NETWORKS
These are networks that are physical and easier for us to understand but in terms of scientific networks they are abstract and harder to follow.

What is a basic network problem?
Moving flow from one node to another in a way that is as efficient as possible (ex: nobel prizes in transportation)

The study of networks involves:
-how to model such applications as mathematical entities
-how to study the models qualitatively (huge impacts if you add a bridge or destroy a node)
-how to design algorithms to solve the resulting problems

CLASSICAL NETWORKS AND APPLICATIONS
examples of classic network problems:
-shortest path problem (ex: emergency fire exiting, DNA sequence alignment, storing books in libraries)
-maximum flow problem (capacities of network need to be determined ex: building evacuation determining capacities of stairwells)
-minimum cost flow problem (certain costs and flow patterns need to be determined to minimize cost ex: cash management)

She says that this is a point she really wants to get across:

The study of networks is not limited to physical networks but also to abstract networks in which nodes do not coincide to locations in space.

INTERDISCIPLINARY IMPACT OF NETWORKS
(Why this lady thinks the study of networks is important) or
Advantages of scientific network formalism:
Many of our problems have to do with flows. Ex: migration (people), stock market (capital)
It also provides visual depiction of problems with gives the opportunity to map out a problem visually.
This, she claims, is what science is all about unifying methodology. It’s about structuring the world around us to discover patterns and unify diverse applications. This is the idea she has been really pushing by giving such a diversity in examples of different networks thus far.

“Network theory provides us with a powerful methodology to establish connections with different disciplines and to break down boundaries.”

Here are some more examples of fields that are using networks:
Finance: International trade, optimization problems, accounting applications
Sociology: Social Networks, Organizational Theory
Computer Science: Routing algorithms
Biology: DNA Sequencing
Engineering: Energy, Manufacturing, Transportation

CHARACTERISTICS OF NETWORKS TODAY
What is different from the networks of today?
-In the 50′s and 60′s congestion wasn’t that big of an idea and networks weren’t as large scale (internet).
The loss of productivity because of traffic is $150 billion annually in Europe and $100 billion in the US. People in LA spend more time in traffic than on vacation.
-behavior of networks
user optimized behavior or user-optimized means that a traveler chooses their route and the best or most efficient way for them to get from one node (their house) to another (work). This may be the best way from their perspective, but perhaps it’s not the same as a societal one (system-optimization). System optimization means that there is a control over the network. She says that in the free world there is no traffic control as a system optimization. This makes me think about the less-free world. In Colombia there are rules about how often and on what days of the week a person can drive to work to avoid total traffic congestion and force people to sometimes take the public transportation or find an alternative route. This is a systematized optimization traffic network. The United States has a user-optimized one.

THE BRAESS PARADOX
-interactions among networks
Interaction among networks is describes with the Braess Paradox. I’ll do my best to sort this out because she illustrates it using all sorts of big letters, little letters, numbers and parenthesis that generally give me hives and I left behind in my high school short term memory. If you happen to be interested in reading a whole book about the Braess Paradox it’s just been translated into English.
My basic understanding is that if there is a traveler who wants to get to a certain point and they have two routes that are equal that means that there is an equilibrium path within their network. But, if an additional link is places in the travel route it will up the cost because it ruins the equilibrium of the flow. (Slide 45) This additional link is an added option, but the travel on the new path increases all of the travel time for everyone. In New York City on Earth Day in 1990 42nd Street was closed and no one complained and everyone was better off because the time was actually made more productive. In Germany a road was added in a major intersection, giving travelers an alternative, and they hated it and complained so it made the time of travel worse. This is supposed to illustrate how the behavior of a group must be taken into account in networks. This is the paradox of Braess’ Paradox. I’m kind of lost in this one so hopefully Petra’s number knowledge is better than mine.

SUPERNETWORKS
A network like the internet is called a noncooperative network. This calls for a new paradigm, some have called it “the price of anarchy” she chooses to call it a supernetwork.
Supernetworks are composed of networks like transportation, telecommunications, logical or financial.
They can be multileveled like studying the supply chain or multitiered as in financial networks.
The tools that they use in research of supernetworks like the internet include:
-network theory
-optimization theory
-game theory (our favorite)
-variational inequality
-projected dynamical systems theory
-network visualization tools
“We are interested not only in addressing topological issues in terms of connectivity, but in predicting the various flows on the networks whether physical or abstract subject to human decision-making under the associated constraints, be they budget, time, security, risk, and/or cost-related.”
Supernetworks seem to be the merging of what we previously saw as separate networks or entities. For example the internet and shopping (ebay), the internet and communication (AIM), supply chain networks with electronic commerce, financial networks with electronic transactions, reverse supply chains with e-cycling, knowledge networks, energy networks and power grids.
She gives us these visual conceptualizations of how each of these fields are affected by supernetworks (the internet).

NOVEL APPLICATIONS
Social Networks and Supernetworks: “It’s not what you know it’s who you know”
-maximization of profit
-minimization of risk
-maximization of relationship value
What are the flows? Relationships and strength of relationships.
Higher relationship levels reduce transaction costs and reduce risk and increase value.
A supernetwork will have a cybernetwork related to a social network.
If you have higher relationship levels you might have lower transaction costs because of the trust level.
Finance literature is increasingly include social networks because of the effect of relationship levels to help aid maximization of profits.

Final points:
-networks are pervasive
-an understating of the tools
-reality of networks that already exist (traffic congestion)
-application of how networks can span disciplines
-bringing the world closer together through different areas of study

Possibly Relevant Posts:

• Reading Summaries (0) | petra
• Travelogue 3… Potential Baby Steps? (3) | JessRoss
• Can My Computer Be Smart Please? Thanks. (1) | Angela