robertjn64 / adventofcode_2022

Puzzle solutions for advent of code (2022)

Advent of Code 2022

RobertJN64 Solutions

All solutions were written with no assistance from forums. Solutions are not necessarily optimal - but many use interesting techniques.

The repo also contains the template file I used, as well as the scripts to launch each part quickly.

Summary:

• All puzzles were completed in under 24 hours except for day 25 because of family conflicts around Christmas
• Top 100 on Day 21
• Top 500 many days, top 1000 most days
• Day 19 - neither part solved - solution not fast enough
• Day 16 - only part A - solution not fast enough
• Day 17 - only part A - too lazy to figure out the caching
• Day 22 - only part A - too lazy to work out the 3D rotations
• Day 25 is only one part

Thoughts on Puzzles:

• 1 - 6 - good introductory puzzles
• 7 (storage space) - recursion + some interesting data structures
• 8 (tree heights) confusing wording (or I was tired)
• 9 (ropes) - tricky but fun
• 10 (CRT display) - off by one errors...
• 11 (Monkeys) - first real optimization with LCM
• 12 (Maze) - standard maze solve alg
• 13 (Sorting) - fun with python magic methods leads to top 150!
• 14, 23 - both decent simulation problems, no real optimization needed
• 15 (Beacons) - First serious optimization problem, I missed the best trick but my code finds all possible solutions
• 16 (Valves) - First brutal problem, earlier than last year. Improved my part A solution from 100 to 0.01 seconds, but still not fast enough for part B
• 17 (Tetris) - Needed caching, too lazy to write it
• 18 (Surface Area) - Over optimized this one after the previous 2 days
• 19 (Blueprints) - Brutally slow - I didn't figure out the correct pruning method
• 20 (Mixing Numbers) - Wrote a "optimal" solution that was slower than python std lib
• 21 (Monkey Math) [Top 100] - fun with python eval
• 22 (Cube Maze) - Best puzzle IMO - too bad the optimal solution was hardcoding
• 24 (Blizzard Maze) - Fun maze solver with some cool optimization tricks
• 25 (-=012 Numbers) - First part was easy, generating the numbers was hard because I'm bad at math apparently

Overall - better score than AoC 2021 (44 instead of 43). I doubt I'll come back to any of the puzzles, maybe 19a to optimize or 22b to finish. To improve from here, I need to learn a few more dynamic programming techniques, and get more proficient with parsing in C++ for the slow days.