5 comments

  • threebodyorbits 4 days ago ago

    At this website, users can contribute compute towards finding new periodic solutions to the three body problem via their browsers. Out of nearly 20 billion orbits tested so far, ~43k returned to their starting positions. The hit rate is ~0.0002%. The discovered orbits can be explored at https://www.threebodyorbits.com/hunt/garden

  • Gravityloss 15 hours ago ago

    Thinking about real astronomical three bodies, I assume many of these would not be possible because of too high tidal forces, preventing the bodies from ever becoming solid, or breaking up if forming by capture.

    So perhaps a feature for searching by limiting tidal forces? Don't know what other limits there are making some of these more and some less likely to exist in the real world?

  • don_esteban 14 hours ago ago

    How does this handle the limits of computational precision?

    The way I understand it, three body problem is quite chaotic (small changes can have over time extreme consequences).

    Hence, computing at finite precision might produce a 'cyclic solution' to the three body problem that is in physical reality (ha, assuming the gravity is truly contiguous with infinite precision) not really cyclic (even excluding things like tides).

    • dreamcompiler 10 hours ago ago

      Some configurations of the 3BP are not chaotic and thus do not exhibit sensitive dependence. Finding these configurations is difficult, hence this website.

      • don_esteban 4 hours ago ago

        I just don't understand how does this website help find those non-chaotic configurations.

        Unless there is some numerical magic going on, it may well be the case that a configuration is chaotic, but in each computational step it is nudged a bit by imprecise calculation/rounding error to remain on the cyclical track, thus seeming non-chaotic.