任意多次交换两数组元素使得两数组之和的差值最小

思想：注意两数组长度不一定相同。现只能想到穷举法（时间复杂度大约为O(n^3)）：记当前两数组之和的差值为abs(diff)，其中diff=sum_A-sum_B，遍历所有组合(a[i], b[j])，察看abs(diff-2*(a[i]-b[j]))与abs(diff)的大小，若前者更小，则更新diff -= 2*(a[i]-b[j])并swap(a[i], b[j])；若找不到这样的组合，则当前状况为所求的最优解

代码：

int process(int a[], int b[], int n, int m){
	assert(a != NULL);
	assert(b != NULL);
	assert(n > 0);
	assert(m > 0);

	int diff = 0;
	for(int i = 0; i < n; i++) diff += a[i];
	for(int i = 0; i < m; i++) diff -= b[i];  //初始化diff
		

	for(int i = 0; i < n; i++){
		for(int j = 0; j < m; j++){
			if(abs(diff-2*(a[i]-b[j])) < abs(diff)){
				diff = diff-2*(a[i]-b[j]);  //注意先更新diff再交换a[i]和b[j]
				swap(&a[i], &b[j]);
				i = 0, j = 0;  //若diff被更新则从头开始遍历
			}
		}
	}
	return diff;
}