skip to main |
skip to sidebar
Deep Simplicity - Chapter 3
Chaos Out of Order
- Iterative Process - dependent on prior state
- Logistic Equation - can be used to illustrate period-doubling
- Example: how the population of a species changes
- x (next) = B * x (current) * (1 - a)
- B = Birth rate
- x = Population
- a = Premature death rate (prior to mating)
- If B<1: population dies
- If B>1 & B<3: x eventually settles at roughly 0.66 (2/3 of max population)
- If B>3: single attractor state becomes two (period doubles to 2)
- Practically speaking, this results in 1 year of overpopulation, followed by 1 yr of underpopulation, etc - this cycle continues back and forth (2 stable states)
- Robert May - first to conduct tests of behavior of logistic equations
- B=3.4495 leads to 4 states
- B=3.56 leads to 8 states
- B=3.569 leads to 16 states
- B=3.5699 (or higher) leads to infinite states
- Mostly chaotic (infinite) beyond 3.5699, although there are small pockets of order, which eventually lead to chaos, on a smaller scale (self-similar)
- Chaos to order, chaos to order, etc.
- James Yorke & Tien Yien Li - "Period Three Implies Chaos"
- Mitchell Feigenbaum - Period-doubling is not unique to logistic equations; it is product of iterative feedback process (self-referential)
- Feigenbaum's Number - 4.669:1 is the universal ratio of period-doubling in all self-referential systems
Fractals
- Giuseppe Peano - constructed a curve (Peano Curve) that completely fills a plane (w/o ever crossing)
- Pattern is self-similar and infinitely long, but contained in finite area
- Curves (nonintersecting) are 1 dimensional and planes are 2 dimensional
- Mandelbrot - realized that the Peano Curve is somewhere between 1 & 2 dimensions
- George Cantor - Cantor Set
- Produced by iteration; self-similar
- Divide a line by thirds
- Divide the two outer thirds by thirds, repeat, etc.
- Warclaw Sierpinski - Sierpinski Gasket
- Helge Von Koch - Koch Curve
- Infinitely long (although it has end points)
- Koch snowflake
- Koch island
- Richardson realized measurements of coastlines erred by up to 20%
- Due to progressively smaller scales of measurement
- Each segment can be scaled by a factor of 3 to replicate original (self-similar)
Richardson's coastline paradox
No comments:
Post a Comment