918. Maximum Sum Circular Subarray
Problem Description
The task is to find the maximum sum of a non-empty subarray within a given circular integer array, nums
. A circular array means that the array is conceptually connected end-to-end, so the elements wrap around. For example, in such an array, the next element of the last element is the first element of the array, and the previous element of the first element is the last element of the array.
The critical challenge here is accounting for the circular nature of the array, which allows for subarrays that cross the end of the array and start again at the beginning. The maximum sum could come from a subarray in the middle of the array, a subarray that includes elements from both ends of the array, or even the entire array if all elements are positive.
The subarray we are trying to find should not include the same element more than once, meaning we can't wrap around the circular array more than once when selecting our subarray elements.
Intuition
To solve the maximum subarray problem for a circular array, we must consider two cases:
- The maximum sum subarray is similar to the one found in a non-circular array (Kadane's algorithm is useful here).
- The maximum sum subarray is the result of wrapping around the circular array.
For the first case, we use Kadane's algorithm to find the maximum sum subarray that does not wrap around. This is done by iterating through the array while maintaining the maximum sum ending at each index (denoted as f1
in the code), and the maximum sum so far (denoted as s1
).
For the second case, to handle subarrays that wrap, consider that the answer could be the total sum of the array minus the minimum sum subarray. This is akin to "selecting" the rest of the array that doesn't include the minimum sum subarray. To find the minimum sum subarray, we modify Kadane's algorithm to keep track of the minimum sum ending at each index (denoted as f2
), and the minimum sum so far (denoted as s2
).
At the end, the maximum possible sum for the case when the subarray wraps around is computed as the total sum of the array minus the minimum sum subarray (sum(nums) - s2
).
The edge case to consider is when all numbers in the array are negative. In this situation, Kadane's algorithm yields a subarray sum which is the maximum negative number (i.e., the least negative), and thus is the maximum sum we can achieve without wrapping around. Subtracting any subarray would result in a smaller sum, so we return the maximum subarray sum found without wrapping in this case.
Putting it all together, the final answer is the larger of the two possibilities: the maximum subarray sum found without wrapping (using Kadane's algorithm) or the total array sum minus the minimum subarray sum, unless all numbers are negative, in which case we return the maximum subarray sum without wrapping.
Learn more about Queue, Divide and Conquer, Dynamic Programming and Monotonic Queue patterns.
Solution Approach
The solution approach involves implementing two variations of Kadane's algorithm once for finding the maximum subarray sum, and another time for finding the minimum subarray sum. We then combine the results of these variations to account for the circular nature of the array.
Here's how we approach it step by step:
-
Initialization: First, we initialize four variables
s1
,s2
,f1
, andf2
with the value of the first element in the array.s1
will hold the maximum subarray sum found so far,s2
the minimum subarray sum,f1
the current maximum subarray sum ending at the current index, andf2
the current minimum subarray sum ending at the current index. -
Iterate Over Array: Start iterating the array from the second element because the first element is used for initialization.
-
Kadane’s Algorithm for Maximum Sum: For each element
num
, we calculate the current maximum ending here (f1
) as the maximum ofnum
andnum + f1
(i.e., either start a new subarray from current element or add the current element to the existing subarray). We then update the overall maximum (s1
) with the maximum of itself andf1
. -
Kadane’s Algorithm for Minimum Sum: Likewise, we calculate the current minimum ending here (
f2
) as the minimum ofnum
andnum + f2
(i.e., either start a new subarray from current element or add the current element to the existing negative subarray). We update the overall minimum (s2
) with the minimum of itself andf2
. -
Check for All Negative Elements: After iterating through the array, we check if the maximum subarray sum found (
s1
) is less than or equal to 0. If true, then all the elements are negative, ands1
is our answer as no subarray wrapping can result in a higher sum. -
Find Maximum Sum Considering Circular Wrap: Otherwise, we find the maximum of
s1
andsum(nums) - s2
. This accounts for the possibility that the maximum sum subarray might be wrapping over the circular array. We subtracts2
(the minimum sum subarray) from the total sum of the array, which effectively gives us the sum of the subarray that wraps around. -
Return the Result: The maximum of these two values gives us our answer, which is the maximum sum of a non-empty subarray for the circular array.
The final Python function looks like:
class Solution:
def maxSubarraySumCircular(self, nums: List[int]) -> int:
s1 = s2 = f1 = f2 = nums[0]
for num in nums[1:]:
f1 = num + max(f1, 0)
f2 = num + min(f2, 0)
s1 = max(s1, f1)
s2 = min(s2, f2)
return s1 if s1 <= 0 else max(s1, sum(nums) - s2)
In this code snippet:
max(f1, 0)
andmin(f2, 0)
are used to decide whether to start a new subarray or to continue with the current subarray.sum(nums)
is called only once to improve efficiency, just before the comparison of sums, as it could be computationally expensive for large arrays to call it repeatedly.
Ready to land your dream job?
Unlock your dream job with a 2-minute evaluator for a personalized learning plan!
Start EvaluatorExample Walkthrough
Let's illustrate the solution approach using a small example with the circular array nums = [5, -3, 5]
.
-
Initialization: We initialize
s1
,s2
,f1
, andf2
with the value of the first element in the array. So,s1 = s2 = f1 = f2 = 5
. -
Iterate Over Array: Start iterating from the second element
-3
. -
Kadane’s Algorithm for Maximum Sum:
- For
-3
:f1 = max(-3, 5 - 3) = max(-3, 2) = 2
, sos1 = max(5, 2) = 5
. - For
5
(last element):f1 = max(5, 5 + 2) = max(5, 7) = 7
, which is a wrap-around as we count5
from both ends of the array.s1
is updated tomax(5, 7) = 7
.
- For
-
Kadane’s Algorithm for Minimum Sum:
- For
-3
:f2 = min(-3, 5 - 3) = min(-3, 2) = -3
, sos2
becomesmin(5, -3) = -3
. - For
5
:f2 = min(5, 5 - 3) = min(5, 2) = 2
, ands2
remains at-3
.
- For
-
Check for All Negative Elements: Since the maximum subarray sum
s1
is greater than 0, not all elements are negative. -
Find Maximum Sum Considering Circular Wrap: We find
max(7, sum([5, -3, 5]) - -3)
. The sum ofnums
is7
, and subtractings2
(which is-3
), we get7 - (-3) = 10
. -
Return the Result: The answer is the maximum of
7
and10
, which is10
. So the maximum sum of a non-empty subarray for the given circular arraynums
is10
.
This walkthrough clearly demonstrates how the algorithm incorporates circularity by considering the inverse of the minimum subarray sum to find the potential maximum in a wrapping scenario. It also shows how to handle non-wrap-around scenarios using Kadane’s algorithm directly for the maximum subarray sum.
Solution Implementation
1from typing import List
2
3class Solution:
4 def max_subarray_sum_circular(self, nums: List[int]) -> int:
5 max_sum_end_here = min_sum_end_here = max_subarray_sum = min_subarray_sum = nums[0]
6
7 # Iterate through the given nums list starting from the second element
8 for num in nums[1:]:
9 # Update max_sum_end_here to be the maximum of the current number or the current number plus max_sum_end_here
10 max_sum_end_here = num + max(max_sum_end_here, 0)
11 # Update min_sum_end_here to be the minimum of the current number or the current number plus min_sum_end_here
12 min_sum_end_here = num + min(min_sum_end_here, 0)
13
14 # Update the max_subarray_sum if the newly computed max_sum_end_here is larger
15 max_subarray_sum = max(max_subarray_sum, max_sum_end_here)
16 # Update the min_subarray_sum if the newly computed min_sum_end_here is smaller
17 min_subarray_sum = min(min_subarray_sum, min_sum_end_here)
18
19 # If the max_subarray_sum is non-positive, the whole array could be non-positive
20 # Thus, the max subarray sum is the max_subarray_sum itself
21 if max_subarray_sum <= 0:
22 return max_subarray_sum
23 else:
24 # Otherwise, we compare the max_subarray_sum vs. total_sum minus min_subarray_sum
25 # The latter represents the maximum sum obtained by considering the circular nature of the array
26 # We subtract min_subarray_sum from the total sum to get the maximum sum subarray which wraps around the array
27 total_sum = sum(nums)
28 return max(max_subarray_sum, total_sum - min_subarray_sum)
29
30# Usage example:
31# solution = Solution()
32# print(solution.max_subarray_sum_circular([5, -3, 5])) # Output: 10
33
1class Solution {
2 public int maxSubarraySumCircular(int[] nums) {
3 // Initialize variables to hold
4 // s1 and s2 as the max and min subarray sums respectively
5 // f1 and f2 as the local max and min subarray sums at the current position
6 // total as the sum of all numbers in the array
7 int maxSubarraySum = nums[0], minSubarraySum = nums[0];
8 int currentMaxSum = nums[0], currentMinSum = nums[0], totalSum = nums[0];
9
10 // Iterate through the array starting from the second element
11 for (int i = 1; i < nums.length; ++i) {
12 // Update the total sum
13 totalSum += nums[i];
14
15 // Calculate the local max subarray sum (Kadane's algorithm)
16 currentMaxSum = nums[i] + Math.max(currentMaxSum, 0);
17 // Determine the global max subarray sum so far
18 maxSubarraySum = Math.max(maxSubarraySum, currentMaxSum);
19
20 // Calculate the local min subarray sum (Inverse Kadane's algorithm)
21 currentMinSum = nums[i] + Math.min(currentMinSum, 0);
22 // Determine the global min subarray sum so far
23 minSubarraySum = Math.min(minSubarraySum, currentMinSum);
24 }
25
26 // If maxSubarraySum is non-positive, all numbers are non-positive.
27 // Hence, return maxSubarraySum because wrapping doesn't make sense.
28 // Otherwise, return the maximum between maxSubarraySum and
29 // totalSum - minSubarraySum which represents the maximum circular subarray sum.
30 return maxSubarraySum > 0 ? Math.max(maxSubarraySum, totalSum - minSubarraySum) : maxSubarraySum;
31 }
32}
33
1#include <vector>
2#include <algorithm>
3
4class Solution {
5public:
6 int maxSubarraySumCircular(vector<int>& nums) {
7 int maxEndingHere = nums[0]; // current max subarray ending at this position
8 int minEndingHere = nums[0]; // current min subarray ending at this position
9 int maxSoFar = nums[0]; // max subarray found so far
10 int minSoFar = nums[0]; // min subarray found so far
11 int totalSum = nums[0]; // sum of all elements
12
13 // Loop starting from the second element in nums
14 for (int i = 1; i < nums.size(); ++i) {
15 totalSum += nums[i]; // accumulate the total sum of the array
16
17 // Update maxEndingHere by including the current number or restarting at current number if it is bigger
18 maxEndingHere = nums[i] + std::max(maxEndingHere, 0);
19 // Update maxSoFar to the maximum of itself and maxEndingHere
20 maxSoFar = std::max(maxSoFar, maxEndingHere);
21
22 // Update minEndingHere by including the current number or restarting at current number if it is smaller
23 minEndingHere = nums[i] + std::min(minEndingHere, 0);
24 // Update minSoFar to the minimum of itself and minEndingHere
25 minSoFar = std::min(minSoFar, minEndingHere);
26 }
27
28 // If maxSoFar is less than 0, all numbers are negative, return maxSoFar directly
29 // Otherwise, return the maximum of maxSoFar and totalSum - minSoFar
30 // (since the maximum sum circular subarray may wrap around the end of the array)
31 return maxSoFar > 0 ? std::max(maxSoFar, totalSum - minSoFar) : maxSoFar;
32 }
33};
34
1/**
2 * Finds the maximum sum of a non-empty subarray for a circular array.
3 *
4 * @param {number[]} nums - The input array of numbers.
5 * @returns {number} - The maximum sum of the subarray.
6 */
7function maxSubarraySumCircular(nums: number[]): number {
8 let currentMax = nums[0]; // Initialize current max subarray sum ending at the current position
9 let globalMax = nums[0]; // Initialize the global max subarray sum found so far
10 let currentMin = nums[0]; // Initialize current min subarray sum ending at the current position
11 let globalMin = nums[0]; // Initialize the global min subarray sum found so far
12 let totalSum = nums[0]; // Initialize the total sum of the array
13
14 for (let i = 1; i < nums.length; ++i) {
15 const currentValue = nums[i]; // Current element being processed
16 totalSum += currentValue; // Add the current value to the total sum
17
18 // Calculate max subarray sum for a normal array (not accounting for circularity)
19 currentMax = Math.max(currentMax + currentValue, currentValue);
20 globalMax = Math.max(currentMax, globalMax);
21
22 // Calculate min subarray sum for a normal array (to help with circular cases)
23 currentMin = Math.min(currentMin + currentValue, currentValue);
24 globalMin = Math.min(currentMin, globalMin);
25 }
26
27 // If all numbers are negative, max subarray sum is the maximum element (as taking an empty subarray is not allowed)
28 // Otherwise, return the maximum between the globalMax and totalSum - globalMin (which accounts for wrap-around subarrays)
29 return globalMax > 0 ? Math.max(globalMax, totalSum - globalMin) : globalMax;
30}
31
Time and Space Complexity
The time complexity of the given code is O(n)
, where n
is the length of the input list nums
. This is because the code iterates through all the elements of the array exactly once to calculate the maximum subarray sum for both the non-circular and circular cases.
The space complexity of the code is O(1)
since it only uses a constant amount of extra space for variables s1
, s2
, f1
, f2
, and some temporary variables to perform the calculations, regardless of the size of the input list.
Learn more about how to find time and space complexity quickly using problem constraints.
In a binary min heap, the minimum element can be found in:
Recommended Readings
Queue Intro Think of the last time you stood in line to buy movie tickets The first person in line gets their ticket first and any newcomers join the end of the line This system where the first person to arrive is the first to be served is a queue in real
https algomonster s3 us east 2 amazonaws com cover_photos dnd svg Divide and Conquer The main idea of divide and conquer is to split the main problem into two smaller components assuming that each one of the components had already been solved recursively and then try to solve the bigger
What is Dynamic Programming Prerequisite DFS problems dfs_intro Backtracking problems backtracking Memoization problems memoization_intro Pruning problems backtracking_pruning Dynamic programming is an algorithmic optimization technique that breaks down a complicated problem into smaller overlapping sub problems in a recursive manner and uses solutions to the sub problems to construct a solution
Want a Structured Path to Master System Design Too? Don’t Miss This!