Length of Longest Fibonacci Subsequence — Problem & Solution
Length of Longest Fibonacci Subsequence solution explained: step-by-step approach, complexity and code in Python, Java, C++ and JavaScript. A medium problem.
Length of Longest Fibonacci Subsequence is a medium coding problem on arrays, hash table and dynamic programming, asked in Amazon, Google and Bloomberg interviews.
- Difficulty: Medium
- Topics: Arrays, Hash Table, Dynamic Programming
- Asked at: Amazon, Google, Bloomberg
- Time limit: 2 s
- Memory limit: 256 MB
- Languages: JavaScript, TypeScript, Python, Java, C++, C, C#, Go, Kotlin, Swift, Rust, PHP and Ruby
Problem statement
A sequence x1, x2, ..., xk is Fibonacci-like when k >= 3 and every term from the third on is the sum of the two before it: x(i) + x(i+1) == x(i+2).
arr is strictly increasing. Return the length of the longest Fibonacci-like subsequence of arr (keep elements in their original order, deleting any others), or 0 if there is none.
Example 1
Input: arr = [1,2,3,4,5,6,7,8]
Output: 5
Explanation: `[1,2,3,5,8]`.
Example 2
Input: arr = [1,3,7,11,12,14,18]
Output: 3
Explanation: For example `[1,11,12]` or `[7,11,18]`.
Example 3
Input: arr = [2,4,7]
Output: 0
Constraints
3 <= arr.length <= 10001 <= arr[i] < arr[i + 1] <= 10^9
Hints for Length of Longest Fibonacci Subsequence
Hint 1
A Fibonacci-like sequence is fixed once you know its first two terms — or its last two.
Hint 2
Let
dp[j][k]be the longest such sequence ending witharr[j], arr[k](j < k).Hint 3
The term before
arr[j]must bearr[k] - arr[j]; if it exists at indexi < j(look it up in a hash map), thendp[j][k] = dp[i][j] + 1.
How to solve Length of Longest Fibonacci Subsequence
Index states by the last two elements. Because arr is strictly increasing, the element preceding the pair (arr[j], arr[k]) is forced to be arr[k] - arr[j], and it must be smaller than arr[j].
Approach
- Map every value to its index.
- For every pair
j < k, letp = arr[k] - arr[j]. Ifp < arr[j]andpis in the map at indexi, setdp[j][k] = dp[i][j] + 1, where a missingdp[i][j]counts as 2. - Track the largest
dpvalue produced; it is at least 3 whenever any extension happened. - Return that maximum, or 0 if no pair was ever extended.
Why it works
Every Fibonacci-like subsequence ending in (arr[j], arr[k]) has a unique predecessor term, so the longest one ending in that pair is one longer than the longest ending in (arr[i], arr[j]). Requiring p < arr[j] keeps the indices in order (the array is increasing), so the subsequence order is respected.
Complexity
- Time —
O(n²) - Space —
O(n²)
Pitfalls
- Return 0, not 2, when no three elements form a Fibonacci-like triple.
- Without the
p < arr[j]check you could use an element that comes afterarr[j]. - Values reach 10^9 — compute the difference
arr[k] - arr[j], never the sum of two values in 32-bit arithmetic.
Reference solution
Python
from typing import List
def lenLongestFibSubseq(arr: List[int]) -> int:
n = len(arr)
pos = {v: i for i, v in enumerate(arr)}
dp = [[2] * n for _ in range(n)]
best = 0
for k in range(n):
for j in range(k):
p = arr[k] - arr[j]
if p < arr[j] and p in pos:
i = pos[p]
dp[j][k] = dp[i][j] + 1
if dp[j][k] > best:
best = dp[j][k]
return bestJava
public static int lenLongestFibSubseq(int[] arr) {
int n = arr.length;
Map<Integer, Integer> pos = new HashMap<>();
for (int i = 0; i < n; i++) pos.put(arr[i], i);
int[][] dp = new int[n][n];
int best = 0;
for (int k = 0; k < n; k++) {
for (int j = 0; j < k; j++) {
dp[j][k] = 2;
int p = arr[k] - arr[j];
if (p < arr[j]) {
Integer i = pos.get(p);
if (i != null) {
dp[j][k] = dp[i][j] + 1;
if (dp[j][k] > best) best = dp[j][k];
}
}
}
}
return best;
}C++
int lenLongestFibSubseq(vector<int>& arr) {
int n = arr.size();
unordered_map<int, int> pos;
for (int i = 0; i < n; i++) pos[arr[i]] = i;
vector<vector<int>> dp(n, vector<int>(n, 2));
int best = 0;
for (int k = 0; k < n; k++) {
for (int j = 0; j < k; j++) {
int p = arr[k] - arr[j];
if (p < arr[j]) {
auto it = pos.find(p);
if (it != pos.end()) {
dp[j][k] = dp[it->second][j] + 1;
best = max(best, dp[j][k]);
}
}
}
}
return best;
}JavaScript
var lenLongestFibSubseq = function(arr) {
var n = arr.length;
var pos = new Map();
for (var i = 0; i < n; i++) pos.set(arr[i], i);
var dp = [];
for (i = 0; i < n; i++) dp.push(new Array(n).fill(2));
var best = 0;
for (var k = 0; k < n; k++) {
for (var j = 0; j < k; j++) {
var p = arr[k] - arr[j];
if (p < arr[j] && pos.has(p)) {
dp[j][k] = dp[pos.get(p)][j] + 1;
if (dp[j][k] > best) best = dp[j][k];
}
}
}
return best;
};Also on the editorial tab: C, C#, Go, Kotlin, PHP, Ruby, Rust, Swift, TypeScript.
All 1199 arrays problems · the whole catalogue
Learn the technique: Arrays · Hashing: Hash Maps and Hash Sets