How would you organise a cell of people?

April 16, 2008 3:10 PM

Affinity groups and terrorist cells seem do well by them, but what possible combinations are there of autonomous organisation, and what are their main characteristics? What different ways are there to organise people according to a cell model? I'm interested in an overview of different organisational structures and ways of calculating their characteristics.

I was reading Moon is a harsh mistress and although I can't visualise the system Manuel proposes in the book, it got me thinking of it as a mathematical problem.

The goals of any organisation will differ, but the roles of people should be possible to categorise - both in terms of their function as well as how trusted they are. What is the best way to plan ahead for growth without compromising function, decision making, and security? Internal strife and a resulting hostile takeover is a greater risk the more decentralised and anonymous the organisation is, so how would that be countered?

If you put a flash mob on one end - three people co-ordinating an unlimited number of others who don't know the controlling ones nor each other - and a three people per cell organisation with anonymous coordination with others and cell-level decision making, on the other end, what lies in between and how would you categorise it?

Most info on the subject that I've found has been of the pop-intelligence variation, talking about how security agencies are infiltrating terrorist organisations or whatnot, so I'm curious what you wonderful, resourceful people can come up with; What have been the pros and cons of alternative organisations you have been a part of? What experience have you had of grouping people according to personality traits (Lotsa group dynamics links at Wikipedia) which sounds like something that might be experimented with in edgy corporate enviroments?

I imagine that some of these ideas could be evaluated using artificial-life software, (A straight-forward 'structural analysis' would be simpler yet) so any suggestions for apps (pref osx) that would lend itself to this would be just dandy.

posted by monocultured to human relations (4 comments total)

You might retrofit Cytoscape to your purposes, whatever they are.
posted by Blazecock Pileon at 4:17 PM on April 16, 2008

You may find some answers in the Social Network Analysis literature. If you are familiar with some of the concepts (e.g., matrix algebra, graph theory) you will have an easier time with it. The classic starting point for the ideas of SNA is Wasserman & Faust (1994) "Social Network Analysis."

Kathleen Carley at Carnegie Mellon wrote a paper [PDF] formalizing cellular networks. I think she may have written some other stuff on the advantages/disadvantages of various structures. Her free software ORA is written in java, but appears to be available in windows and linux only. It has the ability to generate cellular (and other) networks with a range of parameters and then calculate various metrics on them (or other real or artificial networks).

There are plenty of other network analysis tools (e.g., Pajek, UCINet, Net Miner, Network Workbench, and others), some of which (or at least Pajek) allow you to create random networks with various attributes. ORA is nice because it has prefab reports designed to identify key actors or strengths of orgs. I know that Carley has some sort of agent-based modeling tool to test networks as well...
posted by i love cheese at 4:47 PM on April 16, 2008

i love cheese gave you some great suggestions above. Here are a few more:

Uncloaking Terrorist Networks uses the software and network theory that i love cheese mentions to analyize the 9/11 hijackers and their relationships.

Starfish and the Spider looks at decentralized organizations ("starfish") compared with top-down organizations ("spiders"). One of the authors, Rod Beckstrom, just became the cyber-security czar at Department of Homeland Security. You can read the intro and first chapter online on the website, and it gets into a lot of the questions you're asking, using examples you're very familiar with: Napster, Skype, Craigslist, etc.

Also check out Marc Sageman's Leaderless Jihad and Understanding Terror Networks.

Wikipedia has a good entry on clandestine cell organization and why it works.

If you're looking for expert researchers, search for papers by Rob Cross at UVA, Andrew Parker at Stanford, Steve Borgatti (who created UCINet) who was at Boston College and just went to Kentucky, and Ron Burt at Chicago. On the corporate side, Kate Ehrlich at IBM does a lot of really, really great work in this area, especially as it relates to organizing corporate structure.

Two organizations you might also want to look at are The Network Roundtable at the University of Virginia (run by Rob Cross), and the International Network for Social Network Analysis.
posted by NotMyselfRightNow at 6:00 AM on April 17, 2008

There is a port of ORA to OSX, though not the latest version. It's not listed on the main download page, but if you fill out the registration & get to the actual download page, you'll find a link to get version 1.6.9 for OSX. There's a number of other SNA packages that run under OSX, here's a few of them: Network Workbench, NetLogo, visone & JUNG.
posted by scalefree at 10:39 AM on April 17, 2008