leetcode
leetcode 2401 ~ 2450
将三个组排序

将三个组排序

难度:

标签:

题目描述

You are given a 0-indexed integer array nums of length n.

The numbers from 0 to n - 1 are divided into three groups numbered from 1 to 3, where number i belongs to group nums[i]. Notice that some groups may be empty.

You are allowed to perform this operation any number of times:

  • Pick number x and change its group. More formally, change nums[x] to any number from 1 to 3.

A new array res is constructed using the following procedure:

  1. Sort the numbers in each group independently.
  2. Append the elements of groups 1, 2, and 3 to res in this order.

Array nums is called a beautiful array if the constructed array res is sorted in non-decreasing order.

Return the minimum number of operations to make nums a beautiful array.

 

Example 1:

Input: nums = [2,1,3,2,1]
Output: 3
Explanation: It's optimal to perform three operations:
1. change nums[0] to 1.
2. change nums[2] to 1.
3. change nums[3] to 1.
After performing the operations and sorting the numbers in each group, group 1 becomes equal to [0,1,2,3,4] and group 2 and group 3 become empty. Hence, res is equal to [0,1,2,3,4] which is sorted in non-decreasing order.
It can be proven that there is no valid sequence of less than three operations.

Example 2:

Input: nums = [1,3,2,1,3,3]
Output: 2
Explanation: It's optimal to perform two operations:
1. change nums[1] to 1.
2. change nums[2] to 1.
After performing the operations and sorting the numbers in each group, group 1 becomes equal to [0,1,2,3], group 2 becomes empty, and group 3 becomes equal to [4,5]. Hence, res is equal to [0,1,2,3,4,5] which is sorted in non-decreasing order.
It can be proven that there is no valid sequence of less than two operations.

Example 3:

Input: nums = [2,2,2,2,3,3]
Output: 0
Explanation: It's optimal to not perform operations.
After sorting the numbers in each group, group 1 becomes empty, group 2 becomes equal to [0,1,2,3] and group 3 becomes equal to [4,5]. Hence, res is equal to [0,1,2,3,4,5] which is sorted in non-decreasing order.

 

Constraints:

  • 1 <= nums.length <= 100
  • 1 <= nums[i] <= 3

代码结果

运行时间: 38 ms, 内存: 16.1 MB


// Java Stream solution
// 思路:
// 1. 使用stream计算每个组的元素数量。
// 2. 计算最小操作次数。
// 3. 使用stream简化代码。

import java.util.Arrays;

public class Solution {
    public int minSteps(int[] nums) {
        long count1 = Arrays.stream(nums).filter(num -> num == 1).count();
        long count2 = Arrays.stream(nums).filter(num -> num == 2).count();
        long count3 = Arrays.stream(nums).filter(num -> num == 3).count();
        long maxCount = Math.max(count1, Math.max(count2, count3));
        return (int)(nums.length - maxCount);
    }
}

解释

方法:

该题解采用动态规划来解决问题。定义一个dp数组,其中dp[j]表示将数组nums的前i个元素调整成j+1组的最小操作次数。对于每个元素nums[i],根据它当前的值,更新下一个状态nxt。具体地,若nums[i]为1,则将其保持在组1是不需要操作的,而移动到组2或组3则至少需要一次操作。类似的逻辑应用于当nums[i]值为2或3时。通过遍历整个数组,不断更新dp数组,最后返回dp数组中的最小值,即为将整个数组调整到最终状态的最小操作次数。

时间复杂度:

O(n)

空间复杂度:

O(1)

代码细节讲解

🦆
如何确定使用动态规划是解决这个问题的最有效方法?
动态规划是解决该问题的有效方法,因为问题本质上涉及到状态转移和决策优化。每个元素可以属于三个不同的组,我们需要找到一种最少的操作方式,使得整个数组符合最终的组排序。动态规划能够通过逐步构建解决方案,逐个元素地进行决策,并记录下每种决策的代价,从而找到全局最优解。此外,问题的子问题重叠,即前一个元素的最优决策影响到下一个元素的决策,这是动态规划适用的典型场景。
🦆
在动态规划过程中,为什么在处理nums[i]等于1时,改变到组2或组3的操作为`min(dp[0], dp[1]) + 1`和`min(dp) + 1`?这样的更新逻辑是基于什么原则?
这种更新逻辑基于最小化操作次数的原则。当`nums[i]`等于1时,它本身已经属于组1,所以维持在组1的操作次数是0 (`dp[0]`)。如果要变更到组2,可以从当前已经是组1的状态(`dp[0]`)或者已经是组2的状态(`dp[1]`)进行变更,由于变更组需要额外操作,所以取这两者的最小值然后加1。对于变更到组3的操作,因为可以从任意组变更过来,所以我们取三个状态的最小值再加1。这样的逻辑确保了每一步的操作都是基于当前最优决策。
🦆
题解中提到的`nxt = [float('inf')] * 3`在算法中起到什么作用?为什么初始化为无穷大?
在动态规划中,`nxt = [float('inf')] * 3`用于在每个阶段保存更新后的状态,即每个元素处理后的最小操作次数。初始化为无穷大是为了确保在进行状态更新时,任何一个实际的操作次数(都是非负数)都会小于无穷大,这样就可以通过比较保证找到真实的最小操作次数。这种初始化方法是常见的技巧,用于简化代码逻辑并确保更新过程中总是获取到有效的最小值。

相关问题