跳到主要内容

第 224 场周赛

Problem A - 可以形成最大正方形的矩形数目

显然一个矩形能够构成的最大正方形的边长为min(l,w)\min(l,w),找出最大边长然后计数即可。当然也可以一次遍历,同时记录当前最大值的频次,不过时间复杂度是一样的。

  • 时间复杂度O(N)\mathcal{O}(N)
参考代码(Python 3)
class Solution:
def countGoodRectangles(self, rectangles: List[List[int]]) -> int:
maxlen = max(min(l, w) for l, w in rectangles)
return sum(1 for l, w in rectangles if min(l, w) == maxlen)

Problem B - 同积元组

先统计出所有的乘积出现的频次。如果一个乘积出现了kk次,则能够构成4k(k1)4k(k-1)个同积元组。

  • 时间复杂度O(N2)\mathcal{O}(N^2)
  • 空间复杂度O(N2)\mathcal{O}(N^2)
参考代码(Python 3)
class Solution:
def tupleSameProduct(self, nums: List[int]) -> int:
n = len(nums)
cnt = collections.Counter(nums[i] * nums[j] for i in range(n) for j in range(i + 1, n))
return sum(4 * value * (value - 1) for value in cnt.values())

Problem C - 重新排列后的最大子矩阵

首先逐列进行预处理,从最下方的行开始逆序进行动态规划,得到从每一个位置开始向下最长延伸的连续11的段的长度。

接下来逐行处理,将这一行对应的预处理结果逆序排列,然后遍历更新最大值。由于已经逆序排列,所以对第ii行的处理遍历到第jj个元素时对应的最大矩形高度就是f[j][i]f[j][i](为提高性能,预处理的ff数组是按照列优先进行存储的),从而对应的矩形面积为(j+1)f[j][i](j+1)\cdot f[j][i](因为jj是从00开始的)。

  • 时间复杂度O(MNlogN)\mathcal{O}(MN\log N),其中MM是矩阵的行数,NN是矩阵的列数。
  • 空间复杂度O(MN)\mathcal{O}(MN)
参考代码(Python)
class Solution:
def largestSubmatrix(self, matrix: List[List[int]]) -> int:
m = len(matrix)
n = len(matrix[0])
f = [[0] * m for _ in range(n)]
for col in range(n):
for row in range(m - 1, -1, -1):
if matrix[row][col] == 1:
f[col][row] = 1
if row + 1 < m:
f[col][row] += f[col][row + 1]
ans = 0
for row in range(m):
g = [f[col][row] for col in range(n)]
g.sort(reverse=True)
for i in range(n):
ans = max(ans, g[i] * (i + 1))
return ans

Problem D - 猫和老鼠 II

本题的难点在于状态图并非DAG(有向无环图),所以,解题关键在于破环。如何破环?我们有两种方法。

方法一:增加时间维,将状态图变为DAG

  • 时间复杂度O(n3m3)\mathcal{O(n^3m^3)},因为总状态数为n2m2Tn^2m^2T,其中T=2nmT=2nm
参考代码(Python)
class Solution:
def canMouseWin(self, grid: List[str], catJump: int, mouseJump: int) -> bool:
rows = len(grid)
cols = len(grid[0])
print('\n'.join(grid))

def neighbors(pos, step):
r, c = pos
yield (r, c)
for dr, dc in [(-1, 0), (0, -1), (1, 0), (0, 1)]:
for k in range(1, step + 1):
nr = r + k * dr
nc = c + k * dc
if nr < 0 or nr >= rows or nc < 0 or nc >= cols or grid[nr][nc] == '#':
break
yield (nr, nc)

for i in range(rows):
for j in range(cols):
if grid[i][j] == 'C':
cat = (i, j)
if grid[i][j] == 'M':
mouse = (i, j)
if grid[i][j] == 'F':
food = (i, j)

@functools.lru_cache(None)
def solve(mouse, cat, turn):
if mouse == cat or cat == food or turn > 2 * rows * cols:
return False
if mouse == food:
return True
if turn % 2 == 0:
for mouse_nxt in neighbors(mouse, mouseJump):
if solve(mouse_nxt, cat, turn + 1):
return True
return False
else:
for cat_nxt in neighbors(cat, catJump):
if not solve(mouse, cat_nxt, turn + 1):
return False
return True

return solve(mouse, cat, 0)

方法二:使用变形拓扑排序,从确定结果的状态逆向求解

  • 时间复杂度O(V+E)=O(n2m2(m+n))\mathcal{O}(V+E)=\mathcal{O}(n^2m^2(m+n))
参考代码(C++)
const int dx[4] = {-1, 0, 1, 0}, dy[4] = {0, -1, 0, 1}; 

class Solution {
int n;

int encode(int c, int m, int m_move) {
return 2 * (c * n + m) + m_move;
}

tuple<int, int, int> decode(int code) {
return {code / (2 * n), code % (2 * n) / 2, code % 2};
}
public:
bool canMouseWin(vector<string>& grid, int catJump, int mouseJump) {
int rows = grid.size(), cols = grid[0].size();
n = rows * cols;
int cat, mouse, food;
vector<vector<int>> madj(n), cadj(n);
for (int i = 0; i < rows; ++i)
for (int j = 0; j < cols; ++j) {
if (grid[i][j] == '#')
continue;
int u = i * cols + j;
if (grid[i][j] == 'C')
cat = u;
if (grid[i][j] == 'M')
mouse = u;
if (grid[i][j] == 'F')
food = u;
madj[u].emplace_back(u);
cadj[u].emplace_back(u);
for (int k = 0; k < 4; ++k) {
for (int d = 1; d <= max(catJump, mouseJump); ++d) {
int ni = i + dy[k] * d, nj = j + dx[k] * d;
if (ni < 0 || ni >= rows || nj < 0 || nj >= cols || grid[ni][nj] == '#')
break;
int v = ni * cols + nj;
if (d <= catJump)
cadj[u].emplace_back(v);
if (d <= mouseJump)
madj[u].emplace_back(v);
}
}
}

int max_state = n * n * 2;
vector<int> result(max_state), in_degree(max_state);
vector<vector<int>> adj(max_state);
queue<int> q;

auto assign_state = [&](int c, int m, int m_move, int state) {
int code = encode(c, m, m_move);
result[code] = state;
q.emplace(code);
};

for (int i = 0; i < n; ++i) {
if (i != food) {
assign_state(i, i, 1, -1);
assign_state(food, i, 1, -1);
assign_state(i, food, 0, -1);
}
}

for (int i = 0; i < max_state; ++i) {
auto [c, m, m_move] = decode(i);
if (m_move) {
for (int nm : madj[m]) {
int pre = encode(c, nm, 0);
adj[pre].emplace_back(i);
in_degree[i]++;
}
} else {
for (int nc : cadj[c]) {
int pre = encode(nc, m, 1);
adj[pre].emplace_back(i);
in_degree[i]++;
}
}
}

while (!q.empty()) {
int curr = q.front();
q.pop();
auto [c, m, m_move] = decode(curr);
for (int pre : adj[curr]) {
in_degree[pre]--;
if (result[pre] != 0)
continue;
if (result[curr] == -1) {
result[pre] = 1;
q.emplace(pre);
} else if (in_degree[pre] == 0) {
result[pre] = -1;
q.emplace(pre);
}
}
}

return result[encode(cat, mouse, 1)] == 1;
}
};