From: hal@finney.org
Date: Wed Sep 15 1999 - 11:57:42 MDT
David Lubkin, <lubkin@unreasonable.com>, writes:
> Given quadrillions of sensors and lots of computational power, how far in
> advance can we predict severe or extreme weather and seismic events?
>
> How can we divert or dissipate them?
One possibility is chaotic control. Chaotic systems like the weather
can theoretically be controlled using very small power inputs. You do
need a lot of computer power and sensors though. The idea is that you
have a desired outcome, and you run your predictions to see if it is
in the range of possible outcomes from the current state. You continue
to adjust your inputs to try to keep the desired outcome roughly in the
middle of the range of possible outputs.
It is basically the same process used to keep your car in the middle of
the lane. The car's path is like a chaotic process, in that a small
divergence in initial conditions (steering wheel position) leads to a
large change in where the car ends up after a while. But you can keep
it in the road by making tiny adjustments to the steering wheel.
I don't have a clear understanding of what level of sensing and prediction
is necessary to make this work. At a minimum you would need sensors at
the level of the kinds of changes you were going to make, and models
which can work from disturbances at that level up to the level of the
weather changes you want to make.
Hal
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:05:09 MST