Complex Systems Paper(s) of the Day...

In addition to my academic research, I am highly interested in developing practical applications of network and complex adaptive systems theory.  Here are a series of papers from Peter Mucha at UNC that outline some techniques that I think have significant practical applications in finance and business process optimization (i.e.,  supply chain optimization, currency portfolio optimization, etc):

• Dynamic Clustering of Exchange Rates 
• Communities in Networks 
• Community structure in time-dependent, multiscale, and multiplex networks
• Community Structure in Congressional Co-Sponsorship Networks

Many of the above papers focus on the time evolution of network community structure.  While much of the early network literature focuses on static networks, real world networks are dynamic and evolve over time.  Thus practical applications of networks would be most useful if they could capture this dynamic time evolution in a meaningful way.  Following this line of thinking...

If one was interested in the time evolution of community structure in inter-bank networks with an eye towards creating indicators of systemic risk...then there are a couple of issues that would need to be addressed:

1. Develop a robust network data set!  This type of analysis is incredibly data intensive.  You need enough of a network data so that you can take cross-section slices at various time intervals (quarterly, monthly, weekly, etc).  Ideally you would need to take daily slices in order to develop systemic risk indicators that would be useful...
2. Develop a MEANINGFUL algorithm to identify community structure in bank networks.  There are a number of algorithms that have been developed to id community structure.  However, as discussed in Matthew Jackson's book on social networks, it is often difficult to identify exactly these algorithms are capturing (i.e., do the identified communities have any economic justification? Or are they simply a statistical artifact of the data?)
3. What network analysis techniques are must appropriate to use as measures of systemic risk?
4. Last but not least, how do you get buy-in from clients that this type of analysis would be useful?

Someone who could solve the above problems would make quite a bit of money...