• PM_ME_VINTAGE_30S [he/him]@lemmy.sdf.org
    link
    fedilink
    English
    arrow-up
    34
    ·
    edit-2
    11 days ago

    Also, for anyone who wants to learn about the ins and outs of mathematical optimization (what this program does), here is an excellent free resource:

    https://web.stanford.edu/~boyd/cvxbook/

    Also check out Nocedal and Wright for more numerical stuff, and all of Bertsekas’s books for more theory.

    IMO, optimization for power systems is its own subfield with its own quirks, i.e. I am not well-informed about it other than that I know they typically use Newton’s method or something derived from it if you read either of the two linked books. However, optimization is a general technique which should be a part of any economy at scale. I.e. revolutionaries should unironically learn optimization if you have the time to do so.

    • bobs_guns@lemmygrad.ml
      link
      fedilink
      English
      arrow-up
      35
      ·
      11 days ago

      It sounds similar to things like OR-Tools, CP-SAT and Gurobi. These algorithms aren’t just useful for controlling the power grid. You can also use them for Cockshott-style Big Computer Cybersyn economy-wide planning. Or deciding which order to do your errands in, or figuring out how many blackjack hands of exactly 21 you can make out of a certain set of cards.

      TLDR: as a pretty experienced computer toucher this sounds significantly more real than I initially expected from reading the headline.

  • PM_ME_VINTAGE_30S [he/him]@lemmy.sdf.org
    link
    fedilink
    English
    arrow-up
    25
    ·
    edit-2
    11 days ago

    I’ll say exactly what I felt when I used PSSE and fmincon and every other closed-source optimization program for the first time: “Neato, now where’s the source code? Why am I supposed to just trust that this will work every time?” Or at least put out a paper.

    • ☆ Yσɠƚԋσʂ ☆@lemmygrad.mlOP
      link
      fedilink
      English
      arrow-up
      14
      ·
      edit-2
      11 days ago

      An open source solution would be ideal here for sure, not least because these algorithms are likely generally useful for optimizing any dynamic system of stocks and flows.

      • PM_ME_VINTAGE_30S [he/him]@lemmy.sdf.org
        link
        fedilink
        English
        arrow-up
        11
        ·
        11 days ago

        Absolutely. Whether this particular solver is useful or not for a given problem will depend on how closely your problem matches power system dynamics, but I do think a lot of problems fit into the structural assumptions that power systems optimizers make.

  • aaravchen@lemmy.zip
    link
    fedilink
    English
    arrow-up
    12
    ·
    11 days ago

    Was it because the “super code” password was “password”? From what I understand, that’s yow secure the US power grid is.