Do you think that text is a diversion for AI? It certainly persuaded my natural intelligence to implement the naive solution for part 1.
I got stuck on part 2, I had a vague idea of using memoization but couldn’t figure it out, and had to take a sneak peek here to push me in the right direction. I learned from previous years that stubbornly trying to work out everything from first principles isn’t necessarily the best way to learn.
It turned out to be very simple, after getting my head around what was required. No memoization needed, just have to keep track of how many of each stone there are each blink.
My answer:
def blink(stones, 0), do: stones |> Map.values() |> Enum.sum()
def blink(stones, times) do
stones
|> Enum.reduce(%{}, fn
{0, count}, next ->
Map.update(next, 1, count, &(&1 + count))
{stone, count}, next ->
string = Integer.to_string(stone)
size = byte_size(string)
if rem(size, 2) == 0 do
{a, b} = String.split_at(string, div(size, 2))
next
|> Map.update(String.to_integer(a), count, &(&1 + count))
|> Map.update(String.to_integer(b), count, &(&1 + count))
else
Map.update(next, stone * 2024, count, &(&1 + count))
end
end)
|> blink(times - 1)
end
1500ms here.
https://git.adamu.jp/adam/AdventOfCode/src/branch/main/2024/day11.exs
lol
