Find the contiguous subarray within an array (containing at least one number) which has the largest PRoduct.
For example, given the array [2,3,-2,4], the contiguous subarray [2,3] has the largest product = 6. 这道题属于动态规划的题型,之前常见的是Maximum SubArray,现在是Product Subarray,不过思想是一致的。 当然不用动态规划,常规方法也是可以做的,但是时间复杂度过高(TimeOut),像下面这种形式:
// 思路:用两个指针来指向字数组的头尾int maxProduct(int A[], int n){ assert(n > 0); int subArrayProduct = -32768; for (int i = 0; i != n; ++ i) { int nTempProduct = 1; for (int j = i; j != n; ++ j) { if (j == i) nTempProduct = A[i]; else nTempProduct *= A[j]; if (nTempProduct >= subArrayProduct) subArrayProduct = nTempProduct; } } return subArrayProduct;}用动态规划的方法,就是要找到其转移方程式,也叫动态规划的递推式,动态规划的解法无非是维护两个变量,局部最优和全局最优,我们先来看Maximum SubArray的情况,如果遇到负数,相加之后的值肯定比原值小,但可能比当前值大,也可能小,所以,对于相加的情况,只要能够处理局部最大和全局最大之间的关系即可,对此,写出转移方程式如下: local[i + 1] = Max(local[i] + A[i], A[i]);
global[i + 1] = Max(local[i + 1], global[i]);
对应代码如下:
int maxSubArray(int A[], int n){ assert(n > 0); if (n <= 0) return 0; int global = A[0]; int local = A[0]; for(int i = 1; i != n; ++ i) { local = MAX(A[i], local + A[i]); global = MAX(local, global); } return global;}而对于Product Subarray,要考虑到一种特殊情况,即负数和负数相乘:如果前面得到一个较小的负数,和后面一个较大的负数相乘,得到的反而是一个较大的数,如{2,-3,-7},所以,我们在处理乘法的时候,除了需要维护一个局部最大值,同时还要维护一个局部最小值,由此,可以写出如下的转移方程式:
max_copy[i] = max_local[i] max_local[i + 1] = Max(Max(max_local[i] * A[i], A[i]), min_local * A[i])
min_local[i + 1] = Min(Min(max_copy[i] * A[i], A[i]), min_local * A[i])
对应代码如下:
#define MAX(x,y) ((x)>(y)?(x):(y))#define MIN(x,y) ((x)<(y)?(x):(y))int maxProduct1(int A[], int n){ assert(n > 0); if (n <= 0) return 0; if (n == 1) return A[0]; int max_local = A[0]; int min_local = A[0]; int global = A[0]; for (int i = 1; i != n; ++ i) { int max_copy = max_local; max_local = MAX(MAX(A[i] * max_local, A[i]), A[i] * min_local); min_local = MIN(MIN(A[i] * max_copy, A[i]), A[i] * min_local); global = MAX(global, max_local); } return global;}更加优化的代码,空间复杂度降低为O(1)的代码:
int maxProduct(vector<int>& nums) { int len = nums.size(); int maxherepre = nums[0]; int minherepre = nums[0]; int maxsofar = nums[0]; int maxhere, minhere; for(int i = 1; i < len; i++){ maxhere = max(max(maxherepre*nums[i],minherepre*nums[i]),nums[i]); minhere = min(min(maxherepre*nums[i],minherepre*nums[i]),nums[i]); maxsofar= max(maxsofar,maxhere); maxherepre = maxhere; minherepre = minhere; } return maxsofar; }总结:动态规划题最核心的步骤就是要写出其状态转移方程,但是如何写出正确的方程式,需要我们不断的实践并总结才能达到。
新闻热点
疑难解答