题目:
Find the contiguous subarray within an array (containing at least one number) which has the largest sum.
For example, given the array [-2,1,-3,4,-1,2,1,-5,4]
,
the contiguous subarray [4,-1,2,1]
has the largest sum = 6
.
If you have figured out the O(n) solution, try coding another solution using the divide and conquer approach, which is more subtle.
思路:
1、动态规划法:定义dp[i]表示以位置i结尾的子数组的最大和,则递推方程为:dp[i] = max(dp[i-1] + nums[i], nums[i])。也就是说以位置i结尾的子数组的最大和有且仅有两种情况:一种是以位置i-1结尾的子数组的最大和加上元素i;一种是元素i的值(即该最大和子数组仅包含元素i),具体取决于dp[i-1]是否大于0。由于递推方程中dp[i]仅仅依赖于dp[i-1],所以可以将空间复杂度从O(n)优化到O(1),具体见下面的代码片段1。
2、分治法:虽然分治法的时间复杂度要高于动态规划法,但其思想却颇为巧妙:对于数组nums[left, ..., mid, ..., right],其子数组的最大和只有可能出现在如下三种情况下:1)出现在子数组nums[left, mid]中;2)出现在子数组nums[mid + 1, right]中;3)出现在nums[..., mid, mid + 1, ...]中,即同时跨越nums[mid]和nums[mid+1]。那么就可以采用分治法实现,其关键在于合并部分,具体见下面的代码片段2。该方法中T(n) = 2 * T(n/2) + O(n),所以其总体时间复杂度为O(nlogn),空间复杂度仍然为O(1)。
代码:
1、动态规划法:
class Solution {
public:
int maxSubArray(vector<int>& nums) {
if(nums.size() == 0)
return std::numeric_limits<int>::min();
int max_sum = nums[0];
int sum = nums[0];
for(size_t i = 1; i < nums.size(); ++i)
{
sum = std::max(sum + nums[i], nums[i]);
max_sum = max(max_sum, sum);
}
return max_sum;
}
};
2、分治法:
class Solution {
public:
int maxSubArray(vector<int>& nums) {
if(nums.size() == 0)
return std::numeric_limits<int>::min();
return merge(nums, 0, nums.size() - 1);
}
private:
int merge(vector<int>& nums,int s, int t){
if(s == t) return nums[t];
int mid = (s + t) >> 1;
int side_max = max(merge(nums,s,mid),merge(nums,mid + 1,t));
int left_max = nums[mid], right_max = nums[mid + 1];
int cur = left_max;
for(int i = mid - 1;i >= s; i--) // the max sum that ends at nums[mid]
{
cur += nums[i];
left_max = max(cur,left_max);
}
cur = right_max;
for(int i = mid + 2;i <= t;i++) // the max sum that starts at nums[mid + 1]
{
cur += nums[i];
right_max = max(cur,right_max);
}
return max(side_max,left_max + right_max);
}
};