Scott wrote:
> Does
> any of this research support the view that there is a scale of
> constraint and that more constraint implies more stability or
> predictabiliy (as everyone but Bruce seems to think) or less (pace
> Bruce)?
To be precise my position was merely that it is only in special
circumstances that more constraints imply greater predictability. I
hypothesised that this may include situations when you are already
highly constrained.
There is a body of work in computer science on the difficulty of
modelling and solving problems based on how constrained the situation
is. Below is a sample of references.
-------------------------------------------------
The web page: "Phase Transitions in Search" at Xerox Parc
"A major result of this work is that hard instances of ... are
concentrated near an abrupt transition between under- and
overconstrained problems. This transition is analogous to phase
transitions seen in some physical systems."
at URL:
http://www.parc.xerox.com/spl/groups/dynamics/www/constraints.html
This has many links and an introduction to the area.
---------------------------------------------------
A technical CS article:
Barbara Smith
Phase transition and the mushy region in constraint satisfaction
Proceedings of the 11th European Conference on Artificial Intelligence,
pp. 100-104, 1994.
This shows how problems are easy with either little or much constraint.
Most of the hardest problems lie in between on a phase transition
between being over and under constrained.
--------------------------------------------------
An article on the application to scheduling in a job-shop can be found
at:
ftp://fas.sfu.ca/pub/cs/techreports/1997/CMPT97-21.pdf
Here the maximum difficulty of scheduling occurs at a constrainedness of
0.2, so above this increasing the constraints will simplify the
problem. Here is surely a good example where an agent trying to manage
such will not need much cognition when the situation is either over or
under constrained.
----------------------------------------------------------
A discussion more accessible and relevant to social simulation matters.
Arguing how the complexity of a situation is related to the simultaneous
presence of variety and some constraint is:
The Growth of Structural and Functional Complexity
during Evolution by Francis HEYLIGHEN
at: http://pespmc1.vub.ac.be/papers/ComplexityGrowth.html
--------------------------------------------------
Bruce Edmonds,
Centre for Policy Modelling,
Manchester Metropolitan University, Aytoun Bldg.,
Aytoun St., Manchester, M1 3GH. UK.
Tel: +44 161 247 6479 Fax: +44 161 247 6802
http://www.cpm.mmu.ac.uk/~bruce
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|
