7 ms·

I'm confused, we've known about this for the past 60+ years, most commonly as a method for avoiding local minima/maxima when performing simulated annealing. Wha

• andy_ppp 3y ago
  Yes! I said exactly the same thing, it seems very strange the the whole body of work around operations research with heuristics like Tabu search, squeaky wheel and other ways to try to shake out closer to optimal without the expense of enumerating every solution…
  • romusha 3y ago
    Quanta is running out of cool, deep materials
• T-A 3y ago
  From the article: Grimmer’s paper focused only on smooth functions, which have no sharp kinks, and convex functions, which are shaped like a bowl and only have one optimal value at the bottom.