AdventOfCode2021FSharp VS advent-of-code-kotlin-2021

I was having a hard time figuring out day 12 of Advent of Code. In the end I gave up and went looking for some solutions. I came across this one and this one. However, I didn't quite understand what they did.

F# solution in Jupyter Notebook. Got some help from this thread.

F# solution in Jupyter Notebook. Part 1 takes 500ms but Part 2 takes 24s. Any suggestions on where performance could be improved would be appreciated!

F# solution in Jupyter Notebook. Thank God for the Haskell programmers! I would not have been able to solve this without peeking at their some of their code. Parsing the input string was just the first of the challenges! I envy those dynamic programming languages with an eval function.

F# solution with Jupyter Notebook. Brute force. I actually solved Part 2 before solving Part 1, unknowingly.

F# with Jupyter Notebook. Learned how to use recursive types today! Half-way through the problem I was starting to wonder whether BITS was going to be the new IntCode.

So this year I'm learning F# and functional programming. I could not figure out a immutable solution. So I went with an imperative, mutable solution and did pretty much a straight port from Red Blob Games including using a PriorityQueue.

F# with Jupyter Notebook.

F# with Jupyter Notebook. Slow but works. I'll probably come back to this day and optimize the solution.

F# in Jupyter Notebook. Decided to go with an infinite sequence (generator pattern) for the steps. This made Part 2 a breeze!

Kotlin - I went with solving it in place ( sort of, I had to replace [ ] with ( ) because of the way I stored the numbers ). My first day attempt of it I tried to do it with binary tree, but I'm just not familiar enough with the subject to do it that way and I failed miserably ( I'm studying the subject now, as I think it will be helpful with future code challenges ). So my next day attempt I figured I'd try solving it in place and everything worked out really well ( solved it yesterday, but wanted today to review my code before submitting it ).

Kotlin - it's probably a bit of a memory hog compared to other solutions, but I didn't want to have to deal with index numbers anymore than I had to.

Kotlin - I didn't finish it until three in the morning and just now got around to cleaning up the code. I really struggled with part 2 ( path finding was one of my least favorite subject in computer science ). Though it probably wouldn't have taken so long if I had just scrapped part 2 when it became a mess, instead of wasting time trying to fix it.

Kotlin solution. I converted them to numbers ( like round brackets would be 1 and -1 ), and then just used the numbers to work with.

I think my solution is a binary search. I originally went with a brute force implementation, but when I started the second part it occurred to me that I could do a binary search that would reduce the number of indexes I had to check the cost of moving to.