有一个整数数组{1,2,3,-1,-3,2,5},我的工作是打印导致子数组最大和的元素,得到的和是通过加数组中不相邻的元素。
我使用动态编程编写了给出最大和的代码。但无法打印元素。
public static int maxSum(int arr[])
{
int excl = 0;
int incl = arr[0];
for (int i = 1; i < arr.length; i++)
{
int temp = incl;
incl = Math.max(Math.max(excl + arr[i], arr[i]), incl);
excl = temp;
}
return incl;
}
示例:
{1,2,3,-1,-3,2,5}
应该返回{1,3,5}
,因为最大金额为 9 {4,5,4,3}
有两个和{4,4}
和{5,3}
,在对两个数组进行排序时,我们得到{4,4}
和{3,5}
,因为3 <4我们打印了{{1 }}。(包含第一个最小元素的数组)。答案 0 :(得分:6)
您可以保留一个数组来跟踪用于index of elements
的{{1}}。
我已经使用父数组修改了您的代码以跟踪该代码。另外,根据我的理解,我更改了一些变量名。
add to the current element
我相信可以对代码进行很多清理,但这是以后的工作:)
输出:
public static void maxSum(int[] arr){
int n = arr.length;
int[] parent = new int[n];
parent[0] = -1;
int lastSum = 0; // last sum encountered
int lastPos = -1; // position of that last sum
int currSum = arr[0]; // current sum
int currPos = 0; // position of the current sum
for (int i = 1; i < n; i++) {
parent[i] = lastPos; // save the last sum's position for this element
// below this it is mostly similar to what you have done;
// just keeping track of position too.
int probableSum = Integer.max(arr[i] + lastSum, arr[i]);
int tSum = currSum;
int tPos = currPos;
if(probableSum > currSum){
currSum = probableSum;
currPos = i;
}
lastSum = tSum;
lastPos = tPos;
}
System.out.println(currSum); // print sum
System.out.println(Arrays.toString(parent)); // print parent array; for debugging purposes.
// logic to print the elements
int p = parent[n - 1];
System.out.print(arr[n - 1] + " ");
while (p != -1) {
System.out.print(arr[p] + " ");
p = parent[p];
}
}
{1,2,3,-1,-3,2,5} => 5 3 1
更新。添加了一些代码说明
{4,5,4,3} => 3 5
和lastSum
的值在循环执行期间发生变化。通过观察其值在循环内如何变化,可以最好地理解它们。
在循环currSum
的第i
个迭代的开始期间,拥有可以添加到第lastSum
个元素的最大值;因此基本上可以通过迭代第i
个元素来获得最大值。
i-2
拥有可通过迭代第currSum
个元素而获得的最大值。
在循环结束时,将i-1
添加到第lastSum
元素,并指定为i
。如果currSum
小于0,则将第lastSum
个元素本身指定为i
。现在currSum
的旧值称为currSum
lastSum
和lastPos
拥有各自总和值的最新索引。
在下面显示的每个迭代的所有状态下,最右边的和表示迭代开始时的currPos
。 currSum
左侧的值表示currSum
。它们的索引位置分别记录在lastSum
和currPos
中。
lastPos
保留所使用的par[]
的最后一个索引的值。此数组随后用于构造形成最大不相邻总和的实际元素集。
lastSum
initially
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1
par = -1
i=1 iteration state
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, ?
par = -1, !
// before update
currSum = 1, currPos = 0
lastSum = 0, lastPos = -1
// updating
par[1] = lastPos = -1
probableSum = max(2 + 0, 2) = 2 // max(arr[i] + lastSum, arr[i])
? = max(1, 2) = 2 // max(currSum, probableSum)
! = i = 1
// after update
lastSum = currSum = 1
lastPos = currPos = 0
currSum = ? = 2
currPos = ! = 1
i=2 iteration state
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, 2 ?
par = -1, -1 !
// before update
currSum = 2, currPos = 1
lastSum = 1, lastPos = 0
// updating
par[2] = lastPos = 0
probableSum = max(3 + 1, 3) = 4 // max(arr[i] + lastSum, arr[i])
? = max(2, 4) = 4 // max(currSum, probableSum)
! = i = 2
// after update
lastSum = currSum = 2
lastPos = currPos = 1
currSum = ? = 4
currPos = ! = 2
i = 3 iteration state
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, 2 4 ?
par = -1, -1 0 !
// before update
currSum = 4, currPos = 2
lastSum = 2, lastPos = 1
//updating
par[3] = lastpos = 1
probableSum = max(-1 + 2, -1) = 1 // max(arr[i] + lastSum, arr[i])
? = max(4, 1) = 4 // max(currSum, probableSum) ; no update in ?'s value
! = currPos = 2 // as ?'s value didn't update
// after update
lastSum = currSum = 4
lastPos = currPos = 2
currSum = ? = 4
currPos = ! = 2
i = 4 iteration
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, 2 4 4 ?
par = -1, -1 0 1 !
// before update
currSum = 4, currPos = 2
lastSum = 4, lastPos = 2
// updating
par[4] = lastPos = 2
probableSum = max(-3 + 4, -3) = 1 // max(arr[i] + lastSum, arr[i])
? = max(4, 1) = 4 // max(currSum, probableSum) ; no update in ?'s value
! = currPos = 2 // as ?'s value didn't update
// after update
lastSum = currSum = 4
lastPos = currPos = 2
currPos = ? = 4
currPos = ! = 2
i = 5 iteration
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, 2 4 4 4 ?
par = -1, -1 0 1 2 !
// before update
currSum = 4, currPos = 2
lastSum = 4, lastPos = 2
// updating
par[5] = lastPos = 2
probableSum = max(2 + 4, 2) = 6 // max(arr[i] + lastSum, arr[i])
? = max(4, 6) = 6 // max(currSum, probableSum)
! = i = 5
// after update
lastSum = currSum = 4
lastPos = currPos = 2
currPos = ? = 6
currPos = ! = 5
i = 6 iteration state
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, 2 4 4 4 6 ?
par = -1, -1 0 1 2 2 !
// before update
currSum = 6, currPos = 5
lastSum = 4, lastPos = 2
// updating
par[6] = lastPos = 2
probableSum = max(5 + 4, 5) = 9 // max(arr[i] + lastSum, arr[i])
? = max(6, 9) = 9 // max(currSum, probableSum)
! = i = 6
// after update
lastSum = currSum = 6
lastPos = currPos = 5
currPos = ? = 9
currPos = ! = 6
通过使用par []并循环直到par [p]!= -1,我们可以获得元素索引,这些元素实际上构成了实际所需元素的集合。直接检查代码。
例如
after all iteration state
idx = -1, 0, 1, 2, 3, 4, 5, 6
arr = 0, 1, 2, 3, -1, -3, 2, 5
sum = 0 1, 2 4 4 4 6 9
par = -1, -1 0 1 2 2 2
答案 1 :(得分:5)
我更喜欢酋长的解决方案,但这是另一种方法:
/** returns a list of indices which contain the elements that make up the max sum */
public static List<Integer> maxSum(int arr[]) {
int priorMaxSum = 0;
List<Integer> priorMaxSumList = new ArrayList<>();
// initial max sum
int maxSum = arr[0];
List<Integer> maxSumList = new ArrayList<>();
maxSumList.add(0);
for (int i = 1; i < arr.length; i++) {
final int currSum;
final List<Integer> currSumList;
if (priorMaxSum > 0) {
// if the prior sum was positive, then continue from it
currSum = priorMaxSum + arr[i];
currSumList = new ArrayList<>(priorMaxSumList);
} else {
// if the prior sum was not positive, then throw it out and start new
currSum = arr[i];
currSumList = new ArrayList<>();
}
currSumList.add(i);
// update prior max sum and list
priorMaxSum = maxSum;
priorMaxSumList = new ArrayList<>(maxSumList);
if (currSum > maxSum) {
// update max sum
maxSum = currSum;
maxSumList = currSumList;
}
}
return maxSumList;
}
public static void main(String[] args) throws Exception {
int[] a = {1, 2, 3, -5, -3, 2, 5};
List<Integer> l = maxSum(a);
System.out.println(
"indices {" + l.stream().map(String::valueOf).collect(Collectors.joining(", ")) + "}");
System.out.println("values {"
+ l.stream().map(i -> String.valueOf(a[i])).collect(Collectors.joining(", ")) + "}");
}