thor/advent-of-code-go
1
0
Fork 0
mirror of https://github.com/Threnklyn/advent-of-code-go.git synced 2026-05-18 19:13:27 +02:00
Code Issues Packages Projects Releases Wiki Activity

day18 part2 solution - not perfect but it passes :/

Browse Source
This commit is contained in:
Alex Chao
2020-08-06 19:51:52 -04:00
parent 59a0095488
commit d94081fbd9
3 changed files with 392 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+1 -1
View File
@@ -24,4 +24,4 @@ Day | Name | Type of Algo & Notes
15 | Oxygen System | YAY INTCODE 🙄 <br> - Combination of __2D grid searching algo__, __backtracking algo__ and the the Intcode... <br> - I've realized that I really need to stop using x and y for 2D grids and start using row and col because mathematically x is horizontal and y is vertical... My brain is all jumbled up <br> - Created a Robot struct/class that has a computer inside of it. It goes and searches around, collecting data on the floor types at various coordinates. That data is transformed into a 2D grid/array, and then finally fed into a backtracking, searching algorithm to determine the shortest path (turns out there's only one path to the O2 tank...) <br> - part2 is fairly straight forward 2D grid traversing and tagging a spread of oxygen to valid tiles/hallway spaces
16 | Flawed Frequency Transmission | - Some really weird, contrived (as if this whole thing isn't) phase calculator?.. <br> - part2 broke my brain. Optimally calculate subsets by __precalculating running sums__ (linear), then getting subsets by subtracting two of the precalculated running sums. i.e. sub[2:4] = sub[0:4] - sub[0:2] <br> - And also switching pattern recognition to make calculating a particular output O(nlogn)...
17 | Set and Forget | More Intcode... <br> Robot walking around a scaffolding... Fairly similar to previous algos, and a 2D grid traversal again <br> - I feel like I cheated part2 by finding the pattern by eye after printing what the 2D grid looks like. Then it was a matter of giving the intcode computer the corresponding inputs (in a weird format). <br> - Good example of __iterative approaches being better than recursive approaches__ because the recursive approach in "continuous video mode" causes a stack overflow very quickly
18 | Many-Worlds Interpretation | - Oof, this is the hardest algo so far I think... Combination of __Best Pathfinding Algo__ (going to use a BFS/Dijkstra's) and __Graph Traversal__ <br> - I toyed with the idea of using some kind of bitmap to track keys that were found, but ditched it so I wouldn't have to write my own wrapper around bitwise logic... Time complexity comparisons are roughly similar for comparing two maps as comparing two sets of bits, the space complexity is probably much better for 27-bits (@ and a-z) compared to a hashmap, but OH WELL. <br>
18 | Many-Worlds Interpretation | - Oof, this is the hardest algo so far I think... Combination of __Best Pathfinding Algo__ (going to use a BFS/Dijkstra's) and a weighted __Graph Traversal__ with "locked" edges <br> - I toyed with the idea of using some kind of bitmap to track keys that were found, but ditched it so I wouldn't have to write my own wrapper around bitwise logic... Time complexity comparisons are roughly similar for comparing two maps as comparing two sets of bits, the space complexity is probably much better for 27-bits (@ and a-z) compared to a hashmap, but OH WELL. <br> - The graph traversal required memoizing in order to run in any reasonable amount of time. <br> - My part 2 assumptions would not work for every example which is unfortunate :/ I assumed that within a particular quadrant, if the quadrant does not have the key for a door, just remove the door.