1. 解题思路
这一题的话思路上我是拆成了两步来做的,首先,我们要认识到,这里的变化本质就是看数的二进制表达当中有多少个1,因此,假设给定数字的二进制表示长度为 n n n,我们就是要遍历 1 1 1到 n n n当中有多少数能够在至多 k k k次变换之后变为 1 1 1,显然 k = 1 k=1 k=1时,答案就只有 1 1 1,也就是数字只能包含一个二级位为 1 1 1,然后,对于其他的数,我们只需要用一个迭代遍历即可快速获取,整体的复杂度就是 O ( k n ) O(kn) O(kn)。
然后,剩下的问题就变成了,一直一个数二进制数包含 m m m个 1 1 1,请问一共有多少个不大于给定值 s s s的数,此时我们用一个动态规划即可完成。
2. 代码实现
给出python代码实现如下:
MOD = 10**9+7
factorials = [1 for _ in range(801)]
for i in range(2, 801):
factorials[i] = (i * factorials[i-1]) % MOD
revs = [pow(i, -1, mod=MOD) for i in factorials]
@lru_cache(None)
def C(n, m):
return (factorials[n] * revs[m] * revs[n-m]) % MOD
class Solution:
def countKReducibleNumbers(self, s: str, k: int) -> int:
if s == "1":
return 0
def minus_one(s):
n = len(s)
idx = n-1
while s[idx] == "0":
idx -= 1
s = s[:idx] + "0" + "1" * (n-idx-1)
return s.lstrip("0")
s = minus_one(s)
n = len(s)
digit_nums = defaultdict(set)
digit_nums[1] = {1}
for i in range(2, k+1):
for j in range(2, n+1):
if Counter(bin(j))["1"] in digit_nums[i-1]:
digit_nums[i].add(j)
@lru_cache(None)
def count(idx, ones, allow_large):
if allow_large:
return C(n-idx, ones) if n-idx >= ones else 0
if ones == 0:
return 1
if idx >= n:
return 0
if allow_large:
return (count(idx+1, ones-1, allow_large) + count(idx+1, ones, allow_large)) % MOD
elif s[idx] == "1":
return (count(idx+1, ones-1, allow_large) + count(idx+1, ones, True)) % MOD
else:
return count(idx+1, ones, allow_large)
ans = 0
for digit_num in digit_nums.values():
for digit in digit_num:
ans = (ans + count(0, digit, False)) % MOD
return ans
提交代码评测得到:耗时1513ms,占用内存438.4MB。
标签:Count,digit,return,idx,Less,large,ones,Than,MOD From: https://blog.csdn.net/codename_cys/article/details/143669677