diff --git a/project_euler/problem_345/__init__.py b/project_euler/problem_345/__init__.py new file mode 100644 index 000000000000..e69de29bb2d1 diff --git a/project_euler/problem_345/sol1.py b/project_euler/problem_345/sol1.py new file mode 100644 index 000000000000..642d283ea77a --- /dev/null +++ b/project_euler/problem_345/sol1.py @@ -0,0 +1,93 @@ +""" +Project Euler Problem 345: https://projecteuler.net/problem=345 + +Matrix Sum + +We define the Matrix Sum of a matrix as the maximum possible sum of +matrix elements such that none of the selected elements share the same row or column. +Find the Matrix Sum of the following matrix. + +Brute force solution, with caching intermediate steps to speed up the calculation. + +>>> solution(MATRIX_1) +3315 +""" + +import numpy as np +from numpy.typing import NDArray + +MATRIX_1 = [ + "7 53 183 439 863", + "497 383 563 79 973", + "287 63 343 169 583", + "627 343 773 959 943", + "767 473 103 699 303", +] + +MATRIX_2 = [ + "7 53 183 439 863 497 383 563 79 973 287 63 343 169 583", + "627 343 773 959 943 767 473 103 699 303 957 703 583 639 913", + "447 283 463 29 23 487 463 993 119 883 327 493 423 159 743", + "217 623 3 399 853 407 103 983 89 463 290 516 212 462 350", + "960 376 682 962 300 780 486 502 912 800 250 346 172 812 350", + "870 456 192 162 593 473 915 45 989 873 823 965 425 329 803", + "973 965 905 919 133 673 665 235 509 613 673 815 165 992 326", + "322 148 972 962 286 255 941 541 265 323 925 281 601 95 973", + "445 721 11 525 473 65 511 164 138 672 18 428 154 448 848", + "414 456 310 312 798 104 566 520 302 248 694 976 430 392 198", + "184 829 373 181 631 101 969 613 840 740 778 458 284 760 390", + "821 461 843 513 17 901 711 993 293 157 274 94 192 156 574", + "34 124 4 878 450 476 712 914 838 669 875 299 823 329 699", + "815 559 813 459 522 788 168 586 966 232 308 833 251 631 107", + "813 883 451 509 615 77 281 613 459 205 380 274 302 35 805", +] + + +def solve( + arr: NDArray, row_ind: int, include_set: set[int], cache: dict[str, int] +) -> int: + """ + finds the max sum for array arr starting with row number row_ind, and with columns + included in include_set. cache is used for caching intermediate results. + + >>> solve(np.array([[1, 2], [3,4]]), 0, {0, 1}, {}) + np.int64(5) + """ + + cache_id = f"{row_ind}, {sorted(include_set)}" + if cache_id in cache: + return cache[cache_id] + if row_ind == len(arr): + return 0 + sub_max = 0 + for i in include_set: + new_set = include_set - {i} + sub_max = max( + sub_max, arr[row_ind, i] + solve(arr, row_ind + 1, new_set, cache) + ) + cache[cache_id] = sub_max + return sub_max + + +def solution(matrix_str: list[str] = MATRIX_2) -> int: + """ + Takes list of strings matrix_str to parse the matrix and calculates the max sum. + + >>> solution(["1 2", "3 4"]) + 5 + """ + + n = len(matrix_str) + arr = np.empty((n, n), dtype=np.int64) + for i in range(n): + els = matrix_str[i].strip().split(" ") + for j in range(len(els)): + arr[i, j] = int(els[j]) + + cache: dict[str, int] = {} + ans = solve(arr, 0, set(range(n)), cache) + return int(ans) + + +if __name__ == "__main__": + print(f"{solution() = }")