complex adaptive systems (was special topic COMP473)

This paper is not available except by direct arrangement with me (when it potentially is) and the course number (COMP473) is being used for another reading course, so don't go getting confused now. This would also depend upon the specific approval of school of SMSCS - basically it's only worth pursuing if you're a special case of some kind who simply couldn't live without doing the course. The topics covered below are also possible seeds for Honours research projects.

The Course

Many natural systems, and increasingly many artificial (man-made) systems as well, are characterized by apparently complex behaviors that arise as the result of nonlinear spatio-temporal interactions among a large number of components or subsystems. Examples of such natural systems include immune systems, nervous systems, multicellular organisms, ecologies, and insect societies. Artificial systems sharing this property include parallel and distributed computing systems, large-scale communication networks, artificial neural networks, evolutionary algorithms, large-scale software systems, and economies. Such systems have recently come to be known as Complex Adaptive Systems. 

COMP473 addresses topics in Complex Adaptive Systems. 

Objectives

This course aims to broaden the student's knowledge of some particular aspects of complex adaptive systems. The topics covered are agreed on jointly with the lecturer, and these are then investigated by the student by as much background reading as is possible in the time available. Critical insight into contentious issues, combined with a sense of the wider context, is particularly encouranged. 

Structure

This is in a state of development, but here's an indication: The course begins with 4 weeks of readings, ending in a summary essay. These will set the stage for the later work. A particular area of interest (or perhaps two) within this range will then be decided upon (for each student), and this (these) will be the focus of the rest of the course. Students will hand in an essay proposal, a draft, and the final essay, for each topic. The student and lecturer will meet at each of these points and discuss how things are going individually. The course coordinator is Dr Marcus Frean. The expected workload of the course is 10-12 hours per week. 

Assessment

The details of how this course is to be assessed will be hammered out in conjunction with the students taking it. For example 20% for the initial essay, 20% for the proposal, and 30% each for the draft and final versions (if there is a single focus topic). An oral presentation is also a possibility, to be discussed. There is no exam.

Sample work from previous years

Possible topics for your expansion:

• Ants and swarms - algorithms based on the collective behaviour of insects. See Dorigo (www).

• Gene networks - eg. Dearden and Akam, Nature 406, July 13 2000 (News and Views) and the article to which it refers: von Dassow et al., p 188-192. Jeong, Tombor, Albert, Oltvai and Barabasi, "The large-scale organization of metabolic networks", Nature, 407, 651-654 (2000): Comparative analysis of the networks of 43 organisms. The networks have the same topological scaling properties and show striking similarities to the inherent organization of complex non-biological systems (ie. scale-free random networks). Indicates metabolic organization is identical for all living systems, and that this is due to the robust (error-tolerant) nature of scale-free networks. And Albert, Jeong and Barabasi, "Error and attack tolerance of complex networks", Nature, 406, p378-382 (2000).

• Autocatalytic nets - networks of reactants that achieve "autocatalytic closure", producing all the reactants necessary for their own continuation. eg: "The chicken and egg problem" in John Maynard Smith and Eors Szathmary's book "The major transitions in evolution" (1995), especially section 5.3. Kauffman and/or Dyson.

• Evolutionary and population dynamics -  Lotka-Volterra and Replicator dynamics in models of evolving populations. Perhaps sections from "Evolutionary games and population dynamics", Hofbauer and Sigmund, 1998, and/or Joan Roughgarden's "A primer of ecological theory" (2000).

• Collective dynamics of `small-world' networks. Duncan Watts & Steven Strogatz, Nature, 393, 4 June 1998, p 440. Also "How Robust is the Internet?", Yuhai Tu, Nature 406, 27 July 2000, p353.

• Order for free, and the "edge of chaos" idea. What kinds of complex systems can evolve by accumulation of successive useful variations? Does selection by itself achieve complex systems able to adapt? Are there lawful properties characterizing such complex systems? The overall answer may be that complex systems constructed so that they're on the boundary between order and chaos are those best able to adapt by mutation and selection.Look at Stuart Kauffman's work, eg. "Origins of order" (1993), it's cut-down version "At home in the universe" (1995), or "Investigations" (2000).

• Gaia - James Lovelock put forward the idea in 1979: where's it led to?

• Self-organized criticality. eg. Per Bak's book "How nature works" (1996), and selected papers. See also Levin's book (below). For a popsy personality-based account that's fun if a little old, see also "Complexity", Mitchell Waldrop, 1992.

• Ecologies as complex adaptive systems - eg. Simon Levin's book "Fragile dominion - complexity and the commons" (1999). I'm particularly interested in chapters 5 and 6.

• How adaptation builds complexity - sections of John Holland's 1995 book "Hidden order: how adaptation builds complexity".

• Causality - selected sections of Judea Pearl's book "Causality" (2000) on the role of causality in inference & computation.

• Growing artificial societies - Joshua Epstein and Robert Axtell's 1996 book of the same name (subtitled "Social science form the bottom up"). See also Termites, turtles and traffic jams - Resnick (199?).

• J.Doyne Farmer's work - the second law of organization. "Many of us believe that self-organization is a general property - certainly of the universe, and even more generally of mathematical systems that might be called "complex adaptive systems." Complex adaptive systems have the property that if you run them - by just letting the mathematical variable of "time" go forward - they'll naturally progress from chaotic, disorganized, undifferentiated, independent states to organized, highly differentiated, and highly interdependent states."

• Other topics suggested by the student.