给定一个正元素数组,您必须翻转其中一些元素的符号,以便数组元素的总和应该是最小的非负数(尽可能接近零)。返回最小编号。需要翻转符号的元素的数量,以便结果和是最小的非负数。请注意,所有数组元素的总和不会超过 10 4 。
例子:
Input: arr[] = {15, 10, 6}
Output: 1
Here, we will flip the sign of 15
and the resultant sum will be 1.
Input: arr[] = [14, 10, 4]
Output: 1
Here, we will flip the sign of 14 and the resultant sum will be 0.
Note that flipping the signs of 10 and 4 also gives
the resultant sum 0 but the count of flipped elements is not minimum.
朴素的方法:对于数组的每个元素 A[i], 0 ≤ i < n,我们有 2 个选项。
- A[i] 的符号被翻转(-ve)。
- A[i] 的符号没有翻转(+ve)。
所以我们总共可以有 2 n 种可能的阵列配置。我们可以维护每个配置中元素的总和和翻转次数,并跟踪最小和(平局被最小翻转次数打破)。最小和配置中的翻转次数将是答案。
时间复杂度: O(2 n ),其中 n 是数组中的元素数。
高效方法:这个问题可以使用动态规划解决,是标准 0/1 背包问题的变体。不同之处在于我们有 2 个选项,即要么在背包中包含一个项目,要么排除它,这里就像是否翻转元素的符号。不是背包问题中的袋重,这里是没有翻转的数组所有元素的总和(问题陈述中给出的最大值为 10 4 )。
最优子结构:令 dp[i][j] 是数组的前 i 个元素中所需的最小翻转次数,以使总和等于 j。
1 ≤ i ≤ n 和 -sum ≤ j ≤ sum,其中 sum 是没有翻转的数组所有元素的总和。
If sign of ar[i – 1] is not flipped to make sum = j
dp[i][j] = dp[i – 1][j – A[i – 1]]
If sign of ar[i – 1] is flipped to make sum = j
dp[i][j] = min(dp[i][j], dp[i – 1][j + A[i – 1]]+1)
注意:由于数组元素的总和在翻转后可能为负。所以我们不能使用二维数组进行制表,因为在 dp[i][j] 中,j 是数组的和,索引不能为负。因此,我们将使用哈希映射数组。数组的大小将为 n + 1。
重叠子问题:就像 0/1 背包问题一样,这里也有重叠子问题。我们不需要一次又一次地评估结果,而是可以将子问题的结果存储在表中。
时间复杂度: O(n * sum)
辅助空间: O(n * sum)
其中 n 是元素的数量,sum 是没有翻转的数组元素的总和。
空间优化:如果仔细观察最优子结构,dp[i][j] 将仅取决于 dp[i – 1][j – A[i – 1]]/dp[i – 1][j + A[i – 1]]。因此,只涉及 2 行 i 和 i – 1。因此,我们只需要 2 行而不是 n + 1。
以下是我们优化空间所需的更改。
- 而不是采用数组大小=n+1 声明大小为 2 的数组。
- 引入一个布尔变量(比如标志)来在地图之间切换。我们可以初始化 dp[0] 地图并开始填充 dp[1]。在下一次迭代中,dp[0] 是当前地图,dp[1] 是之前的地图。通过这种方式,我们可以继续在 2 张地图之间切换。
下面是上述方法的实现:
C++
// C++ implementation of the approach
#include
using namespace std;
// Function to return the
// minimum number of elements
// whose sign must be flipped
// to get the positive
// sum of array elements as close
// to 0 as possible
int solve(int A[], int n)
{
// Array of unordered_map of size=2.
unordered_map dp[2];
// boolean variable used for
// toggling between maps
bool flag = 1;
// Calculate the sum of all
// elements of the array
int sum = 0;
for (int i = 0; i < n; i++)
sum += A[i];
// Initializing first map(dp[0])
// with INT_MAX because
// for i=0, there are no elements
// in the array to flip
for (int i = -sum; i <= sum; i++)
dp[0][i] = INT_MAX;
// Base Case
dp[0][0] = 0;
for (int i = 1; i <= n; i++) {
for (int j = -sum; j <= sum; j++) {
dp[flag][j] = INT_MAX;
if (j - A[i - 1] <= sum && j - A[i - 1] >= -sum)
dp[flag][j] = dp[flag ^ 1][j - A[i - 1]];
if (j + A[i - 1] <= sum && j + A[i - 1] >= -sum
&& dp[flag ^ 1][j + A[i - 1]] != INT_MAX)
dp[flag][j]
= min(dp[flag][j],
dp[flag ^ 1][j + A[i - 1]] + 1);
}
// For toggling
flag = flag ^ 1;
}
// Required sum is minimum non-negative
// So, we iterate from i=0 to sum and find
// the first i where dp[flag ^ 1][i] != INT_MAX
for (int i = 0; i <= sum; i++) {
if (dp[flag ^ 1][i] != INT_MAX)
return dp[flag ^ 1][i];
}
// In worst case we will flip max n-1 elements
return n - 1;
}
// Driver code
int main()
{
int arr[] = { 10, 22, 9, 33, 21, 50, 41, 60 };
int n = sizeof(arr) / sizeof(arr[0]);
cout << solve(arr, n);
return 0;
}
Java
// Java implementation of
// the above approach:
class GFG {
// Function to return the
// minimum number of elements
// whose sign must be flipped
// to get the positive
// sum of array elements as close
// to 0 as possible
public static int solve(int[] A, int n)
{
int[][] dp = new int[2000][2000];
// boolean variable used for
// toggling between maps
int flag = 1;
// Calculate the sum of all
// elements of the array
int sum = 0;
for (int i = 0; i < n; i++)
sum += A[i];
// Initializing first map(dp[0])
// with INT_MAX because for i=0,
// there are no elements in the
// array to flip
for (int i = -sum; i <= sum; i++) {
try {
dp[0][i] = Integer.MAX_VALUE;
}
catch (Exception e) {
}
}
// Base Case
dp[0][0] = 0;
for (int i = 1; i <= n; i++) {
for (int j = 0; j <= sum; j++) {
try {
dp[flag][j] = Integer.MAX_VALUE;
if (j - A[i - 1] <= sum
&& j - A[i - 1] >= -sum)
dp[flag][j]
= dp[flag ^ 1][j - A[i - 1]];
if (j + A[i - 1] <= sum
&& j + A[i - 1] >= -sum
&& dp[flag ^ 1][j + A[i - 1]]
!= Integer.MAX_VALUE)
dp[flag][j] = Math.min(
dp[flag][j],
dp[flag ^ 1][j + A[i - 1]] + 1);
}
catch (Exception e) {
}
}
// For toggling
flag = flag ^ 1;
}
// Required sum is minimum non-negative
// So, we iterate from i=0 to sum and find
// the first i where dp[flag ^ 1][i] != INT_MAX
for (int i = 0; i <= sum; i++) {
if (dp[flag ^ 1][i] != Integer.MAX_VALUE)
return dp[flag ^ 1][i];
}
// In worst case we will flip max n-1 elements
return n - 1;
}
// Driver code
public static void main(String[] args)
{
int[] arr = { 10, 22, 9, 33, 21, 50, 41, 60 };
int n = arr.length;
System.out.println(solve(arr, n));
}
}
// This code is contributed by sanjeev2552
Python3
# Python3 implementation of the approach
# Function to return the minimum
# number of elements
# whose sign must be flipped
# to get the positive
# sum of array elements as close
# to 0 as possible
def solve(A, n):
dp = [[0 for i in range(2000)] for i in range(2000)]
# boolean variable used for
# toggling between maps
flag = 1
# Calculate the sum of all
# elements of the array
sum = 0
for i in range(n):
sum += A[i]
# Initializing first map(dp[0])
# with INT_MAX because
# for i=0, there are no elements
# in the array to flip
for i in range(-sum, sum+1):
dp[0][i] = 10**9
# Base Case
dp[0][0] = 0
for i in range(1, n+1):
for j in range(-sum, sum+1):
dp[flag][j] = 10**9
if (j - A[i - 1] <= sum and j - A[i - 1] >= -sum):
dp[flag][j] = dp[flag ^ 1][j - A[i - 1]]
if (j + A[i - 1] <= sum
and j + A[i - 1] >= -sum
and dp[flag ^ 1][j + A[i - 1]] != 10**9):
dp[flag][j] = min(dp[flag][j],
dp[flag ^ 1][j + A[i - 1]] + 1)
# For toggling
flag = flag ^ 1
# Required sum is minimum non-negative
# So, we iterate from i=0 to sum and find
# the first i where dp[flag ^ 1][i] != INT_MAX
for i in range(sum+1):
if (dp[flag ^ 1][i] != 10**9):
return dp[flag ^ 1][i]
# In worst case we will flip max n-1 elements
return n - 1
# Driver code
arr = [10, 22, 9, 33, 21, 50, 41, 60]
n = len(arr)
print(solve(arr, n))
# This code is contributed by mohit kumar 29
C#
// C# implementation of the above approach:
using System;
class GFG
{
// Function to return the minimum number
// of elements whose sign must be flipped
// to get the positive sum of array elements
// as close to 0 as possible
public static int solve(int[] A, int n)
{
int[,] dp = new int[2000, 2000];
// boolean variable used for
// toggling between maps
int flag = 1;
// Calculate the sum of all elements
// of the array
int sum = 0;
for (int i = 0; i < n; i++)
sum += A[i];
// Initializing first map(dp[0]) with
// INT_MAX because for i=0, there are
// no elements in the array to flip
for (int i = -sum; i <= sum; i++)
{
try
{
dp[0, i] = int.MaxValue;
}
catch (Exception e){}
}
// Base Case
dp[0, 0] = 0;
for (int i = 1; i <= n; i++)
{
for (int j = 0; j <= sum; j++)
{
try
{
dp[flag, j] = int.MaxValue;
if (j - A[i - 1] <= sum &&
j - A[i - 1] >= -sum)
dp[flag, j] = dp[flag ^ 1,
j - A[i - 1]];
if (j + A[i - 1] <= sum &&
j + A[i - 1] >= -sum &&
dp[flag ^ 1,
j + A[i - 1]] != int.MaxValue)
dp[flag, j] = Math.Min(dp[flag, j],
dp[flag ^ 1,
j + A[i - 1]] + 1);
} catch (Exception e) {}
}
// For toggling
flag = flag ^ 1;
}
// Required sum is minimum non-negative
// So, we iterate from i=0 to sum and find
// the first i where dp[flag ^ 1,i] != INT_MAX
for (int i = 0; i <= sum; i++)
{
if (dp[flag ^ 1, i] != int.MaxValue)
return dp[flag ^ 1, i];
}
// In worst case we will flip
// max n-1 elements
return n - 1;
}
// Driver code
public static void Main(String[] args)
{
int[] arr = { 10, 22, 9, 33,
21, 50, 41, 60 };
int n = arr.Length;
Console.WriteLine(solve(arr, n));
}
}
// This code is contributed by PrinciRaj1992
Javascript
3
时间复杂度: O(n * sum)。
辅助空间: O(sum) 其中 n 是元素数,sum 是没有翻转的数组元素的总和。
如果您希望与专家一起参加现场课程,请参阅DSA 现场工作专业课程和学生竞争性编程现场课程。