LeetCode 1191 K 次串联后最大子数组之和
贪心
为了方便说明, 定义几个变量。对每个数组arr而言:
maxSub: 最大子数组之和
maxSuf: 最大后缀数组之和
maxPre: 最大前缀数组之和
[opt =
egin{cases}
maxSub &k=1 \
max(maxSuf + s_{arr}*(k-2) + maxPre, maxSub) &k>=2
end{cases} \
s_{arr} = max(0, sum(arr))
]
maxSuf与maxPre均可以通过遍历累加解决
maxSub也是通过遍历选取, 不同在于每次都需要贪心选择局部最大值
[tmpSub_i = egin{cases}
max(0, arr_i) &i=0 \
max(0, arr_i, tmpSub_{i-1} + arr_i) &i>=1
end{cases} \
i in [0, len(arr)) \
maxSub = max(tmpSub_i)
]
伪代码如下：
maxSub, tmpSub = 0, 0
for i, v in arr:
tmpSub = max(0, v, tmpSub + v)
maxSub = max(maxSub, tmpSub)
Python3
from typing import List
class Solution:
def kConcatenationMaxSum(self, arr: List[int], k: int) -> int:
maxSub, tmpSub = 0, 0
maxSuf, tmpSuf = 0, 0
maxPre, s = 0, 0
for i, v in enumerate(arr):
tmpSub = max(0, v, tmpSub + v)
maxSub = max(maxSub, tmpSub)
tmpSuf += arr[-1 - i]
maxSuf = max(maxSuf, tmpSuf)
s += v
maxPre = max(maxPre, s)
if k <= 1:
return maxSub
s = max(0, s)
opt = maxSuf + s * (k-2) + maxPre
return max(opt, maxSub) % (10**9 + 7)