Here is my take. I solved part 2 by detecting cycles. It’s not super fast, finishes in about 7 sec. There’s some room for optimizing, but I didn’t want to uglify the code

I had fun with streams in this one. I also implemented a cycle finding algorithm.

Running the cycle finding algorithm on the infinite stream turned out to be too slow, though, so I cheated a little by pulling out 2000 values to a finite enumerable. That is probably long enough to work for most people’s data sets. The cycle in my data starts at 452 minutes.

It finishes in about 12 minutes.

@sasajuric I liked the elegant way in which you cached the data from the infinite stream and detected the cycle at the same time.

Update: Forgot the link.

I caught up just in time to go on vacation: here’s my solution to Day 18:

https://github.com/JEG2/advent_of_code_2018/blob/master/day_18/settlers_of_the_north_pole.exs