[HN Gopher] Strange Attractors
       ___________________________________________________________________
        
       Strange Attractors
        
       Author : T-A
       Score  : 87 points
       Date   : 2022-11-23 09:51 UTC (13 hours ago)
        
 (HTM) web link (syntopia.github.io)
 (TXT) w3m dump (syntopia.github.io)
        
       | IIAOPSW wrote:
       | Quick intro for those who don't know Chaos theory or what a
       | strange attractor is.
       | 
       | In the most general abstract sense suppose you have a system
       | which is in some state and there is a deterministic rule which
       | updates it to the next state. For now we can just represent the
       | states and transitions as a bunch of circles and arrows
       | connecting them. Draw it on a whiteboard if you must. Since
       | transitions are deterministic, each state only transitions to a
       | single other state. There might be two states that lead to one,
       | but never one that leads to two. In general then, no matter which
       | state you start in, it must either eventually lead to some closed
       | loop of the same states over and over again or it runs off the
       | page in an infinite succession of states. The first option is
       | what we would call an "attractor" and the latter which could be
       | aptly called a "repulsor". There's a third possibility I didn't
       | mention (which is hard to conceptualize drawing things as a 2d
       | diagram of states and transitions). You can have the state run
       | off in an infinite succession of states, but rather than running
       | off the page to infinity, it goes around and around visiting
       | similar states over and over but never technically visiting the
       | same state twice. Its a strange attractor.
       | 
       | Now to be less general. If your system states are some continuous
       | parameter, instead of circles and arrows it makes sense to
       | represent the nth state as x_n and to represent the transition
       | rule to the next state as a difference equation x_{n+1} = f(x_n).
       | If your system has more than one defining parameter, do the same
       | thing but with vector valued quantities. Lastly, if your
       | transitions themselves are continuous rather than discrete,
       | replace the difference equation with a differential equation
       | dx/dt = f(x). No matter which of these you are using to represent
       | system states and define how that system updates, the same basic
       | patterns as before are still the only possibility. The solution
       | to your equation either converges to some fixed point, diverges
       | to infinity (usually exponential growth), or the curve stays
       | bounded within some region of state space getting arbitrarily
       | close to forming a loop but always just missing itself.
        
         | ouid wrote:
         | you've described a computer, it seems to me that you mean to
         | put everything in some kind of normed vector space.
        
         | pineconewarrior wrote:
         | I had no idea what a strange attractor was before reading this
         | and seeing this 3d demo. Between the two, I feel like I
         | understand it quite well now.
         | 
         | Thanks for the writeup!
        
       | briga wrote:
       | Great work on the visualization, looks fantastic! Impressive how
       | far web graphics have advanced.
        
       | q_revert wrote:
       | despite having a phd in nonlinear dynamics, I struggle to wrap my
       | head around most of this stuff... but I always found is deeply
       | fascinating.
       | 
       | In one system I studied we had a sequence of period doublings,
       | which followed one of the feigenbaum constants[1]
       | 
       | seeing this happen in the lab, and subsequently seeing it
       | numerically, and uncovering the analytics behind it was a worthy
       | use of 6 years
       | 
       | --- [1] https://en.wikipedia.org/wiki/Feigenbaum_constants
        
       | daniel-thompson wrote:
       | This reminds me of Chaoscope^. It reached EOL a long time ago and
       | the binaries might have succumbed to bitrot by now, but it was a
       | lot of fun back in the day and produced some beautiful
       | visualizations based on strange attractors.
       | 
       | ^ http://www.chaoscope.org/index.htm
        
       | gatane wrote:
       | Is there a book about this? This isnt covered that deep in
       | Dif.Eq. books...
        
         | rikroots wrote:
         | Not a book, but Paul Bourke's website - Fractals, Chaos, Self-
         | Similarity - has a wealth of information.
         | http://paulbourke.net/fractals/
        
         | lanstin wrote:
         | The Topology of Chaos, https://www.amazon.com/Topology-Chaos-
         | Alice-Stretch-Squeezel...
         | 
         | is a good modern book on it.
         | 
         | It is a fun problem because it stems from continuous diff eq
         | but the answers come from topological analysis around the fixed
         | sets, and the ways fixed sets can change as continuous
         | parameters change.
        
         | peapicker wrote:
         | "Strange Attractors: Creating Patterns in Chaos" by Julien C.
         | Sprott.
         | 
         | I am lucky enough to have a 1st edition print, but there is now
         | a free PDF online: https://sprott.physics.wisc.edu/SA.HTM
        
         | squaredot wrote:
         | Maybe the most famous is the Lorenz attractor, arising from the
         | Lorenz System (this is one of the simulations shown in the
         | website). In Wikipedia you find a description of the system
         | 
         | https://en.wikipedia.org/wiki/Lorenz_system
        
         | bmitc wrote:
         | * _Chaos and Dynamical Systems_ by David Feldman
         | 
         | * _Exploring Chaos: Theory and Experiment_ by Brian Davies
         | 
         | * _Chaos: From Theory to Applications_ by A.A. Tsonis
         | 
         | * _Chaos and Fractals: An Elementary Introduction_ by David
         | Feldman
         | 
         | * _Nonlinear Dynamics And Chaos_ by Tufillaro, Abbott, Reilly
         | 
         | * _Computers, pattern, chaos, and beauty: Graphics from an
         | unseen world_ by Clifford Pickover
         | 
         | * _A First Course In Chaotic Dynamical Systems_ by Robert
         | Devaney
         | 
         | Just some references from my "to read" collection.
        
         | kkylin wrote:
         | James Gleick's "Chaos" remains, in my view, a great popular
         | book on the subject, and manages to convey many of the ideas at
         | a conceptual level without getting technical. And if you enjoy
         | that, you might also like Strogatz's "Sync."
         | 
         | For slightly more technical treatment, Strogatz's Nonlinear
         | Dynamics and Chaos is now a standard text. It isn't terribly
         | technical and is quite well written and (in my view) easy to
         | read for anyone with a background in vector calculus, diff eqs,
         | & perhaps a little bit of linear algebra. (There are now other
         | good options to, some more mathematical and some more
         | application-oriented, but I still think Strogatz is a good
         | place to start.)
        
           | azepoi wrote:
           | Can you expand on the last part? I mean the good other
           | technical books for someone already a bit familiar who wants
           | a deeper dive? Thanks for the pointers
        
           | edbaskerville wrote:
           | Second Strogatz, does an amazing job linking the math to the
           | visual intuition--probably less rigorous than some texts, but
           | the perfect balance for an intro. Took an amazing course
           | based on that book with Charlie Doering, who has sadly
           | passed. He could draw multicolor, qualitatively-correct 3D
           | attractors freehand on a chalkboard while helpfully narrating
           | the system dynamics.
        
       | IdealeZahlen wrote:
       | What kind of ODE solvers are used to simulate chaotic systems?
       | They must be very accurate if even a small error can result in a
       | completely different result.
        
         | j7ake wrote:
         | You can use the standard Euler method with very small delta T.
        
           | pixelpoet wrote:
           | Respectfully disagree; Euler's method is absolutely terrible
           | because it's unconditionally unstable; far better is to use
           | something like Leapfrog or velocity Verlet (both of which
           | have 2nd order accuracy and better stability, for exactly the
           | same number of derivative evaluations). Euler integration is
           | essentially always the wrong tool for the job.
        
             | karpierz wrote:
             | Would RK4 work for something like this, or does it lack
             | some stability properties?
        
               | jordigh wrote:
               | RK4 is about as unstable. The backwards Euler method is
               | baby's first stable ODE solver. If you want to make RK4
               | stable, you have to change it into the implicit RK4
               | method.
               | 
               | However, the Lorenz system isn't stiff, which is usually
               | what you use stable solvers for, nor any of the other
               | systems here, I believe. RK4 should be fine, or even
               | normal Euler with a reasonable step size. The chaos you
               | see in these systems is not due to numerical inaccuracy.
               | They're inherently chaotic. That's what makes chaos a
               | subject worth studying.
        
               | IIAOPSW wrote:
               | If you want to see the folly of RK4, give it something
               | like a ball bouncing and watch as it bounces slightly
               | higher each time. Subtle at first, but trust me its
               | there. The comment above you is right. If you have
               | conservation of energy and want to keep it that way, use
               | verlet.
               | 
               | Edit: before anyone calls me out on this, the same trick
               | also works when there's no discontinuity in the force
               | function vis a vis collision with the floor. Planetary
               | motion also drifts out of simple orbits. I just picked
               | the ball bouncing because its a more amusing visual.
        
       ___________________________________________________________________
       (page generated 2022-11-23 23:01 UTC)