Open In App

Shortest cycle in an undirected unweighted graph

Improve
Improve
Like Article
Like
Save
Share
Report

Given an undirected unweighted graph. The task is to find the length of the shortest cycle in the given graph. If no cycle exists print -1.

Examples:  

Input: 

Output:
Cycle 6 -> 1 -> 5 -> 0 -> 6

Input: 

Output:
Cycle 6 -> 1 -> 2 -> 6 

Prerequisites: Dijkstra 

Approach: For every vertex, we check if it is possible to get the shortest cycle involving this vertex. For every vertex first, push current vertex into the queue and then it’s neighbours and if the vertex which is already visited comes again then the cycle is present. 
Apply the above process for every vertex and get the length of the shortest cycle.

Below is the implementation of the above approach: 

C++




// C++ implementation of the approach
#include <bits/stdc++.h>
using namespace std;
#define N 100200
 
vector<int> gr[N];
 
// Function to add edge
void Add_edge(int x, int y)
{
    gr[x].push_back(y);
    gr[y].push_back(x);
}
 
// Function to find the length of
// the shortest cycle in the graph
int shortest_cycle(int n)
{
    // To store length of the shortest cycle
    int ans = INT_MAX;
 
    // For all vertices
    for (int i = 0; i < n; i++) {
 
        // Make distance maximum
        vector<int> dist(n, (int)(1e9));
 
        // Take a imaginary parent
        vector<int> par(n, -1);
 
        // Distance of source to source is 0
        dist[i] = 0;
        queue<int> q;
 
        // Push the source element
        q.push(i);
 
        // Continue until queue is not empty
        while (!q.empty()) {
 
            // Take the first element
            int x = q.front();
            q.pop();
 
            // Traverse for all it's childs
            for (int child : gr[x]) {
 
                // If it is not visited yet
                if (dist[child] == (int)(1e9)) {
 
                    // Increase distance by 1
                    dist[child] = 1 + dist[x];
 
                    // Change parent
                    par[child] = x;
 
                    // Push into the queue
                    q.push(child);
                }
 
                // If it is already visited
                else if (par[x] != child and par[child] != x)
                    ans = min(ans, dist[x] + dist[child] + 1);
            }
        }
    }
 
    // If graph contains no cycle
    if (ans == INT_MAX)
        return -1;
 
    // If graph contains cycle
    else
        return ans;
}
 
// Driver code
int main()
{
    // Number of vertices
    int n = 7;
 
    // Add edges
    Add_edge(0, 6);
    Add_edge(0, 5);
    Add_edge(5, 1);
    Add_edge(1, 6);
    Add_edge(2, 6);
    Add_edge(2, 3);
    Add_edge(3, 4);
    Add_edge(4, 1);
 
    // Function call
    cout << shortest_cycle(n);
 
    return 0;
}


Java




// Java implementation of the approach
import java.util.*;
 
class GFG
{
 
    static final int N = 100200;
    @SuppressWarnings("unchecked")
    static Vector<Integer>[] gr = new Vector[N];
 
    // Function to add edge
    static void Add_edge(int x, int y)
    {
        gr[x].add(y);
        gr[y].add(x);
    }
 
    // Function to find the length of
    // the shortest cycle in the graph
    static int shortest_cycle(int n)
    {
 
        // To store length of the shortest cycle
        int ans = Integer.MAX_VALUE;
 
        // For all vertices
        for (int i = 0; i < n; i++)
        {
 
            // Make distance maximum
            int[] dist = new int[n];
            Arrays.fill(dist, (int) 1e9);
 
            // Take a imaginary parent
            int[] par = new int[n];
            Arrays.fill(par, -1);
 
            // Distance of source to source is 0
            dist[i] = 0;
            Queue<Integer> q = new LinkedList<>();
 
            // Push the source element
            q.add(i);
 
            // Continue until queue is not empty
            while (!q.isEmpty())
            {
 
                // Take the first element
                int x = q.poll();
 
                // Traverse for all it's childs
                for (int child : gr[x])
                {
                    // If it is not visited yet
                    if (dist[child] == (int) (1e9))
                    {
 
                        // Increase distance by 1
                        dist[child] = 1 + dist[x];
 
                        // Change parent
                        par[child] = x;
 
                        // Push into the queue
                        q.add(child);
                    } else if (par[x] != child && par[child] != x)
                        ans = Math.min(ans, dist[x] + dist[child] + 1);
                }
            }
        }
 
        // If graph contains no cycle
        if (ans == Integer.MAX_VALUE)
            return -1;
 
        // If graph contains cycle
        else
            return ans;
    }
 
    // Driver Code
    public static void main(String[] args)
    {
 
        for (int i = 0; i < N; i++)
            gr[i] = new Vector<>();
 
        // Number of vertices
        int n = 7;
 
        // Add edges
        Add_edge(0, 6);
        Add_edge(0, 5);
        Add_edge(5, 1);
        Add_edge(1, 6);
        Add_edge(2, 6);
        Add_edge(2, 3);
        Add_edge(3, 4);
        Add_edge(4, 1);
 
        // Function call
        System.out.println(shortest_cycle(n));
    }
}
 
// This code is contributed by
// sanjeev2552


Python3




# Python3 implementation of the approach
from sys import maxsize as INT_MAX
from collections import deque
 
N = 100200
 
gr = [0] * N
for i in range(N):
    gr[i] = []
 
# Function to add edge
def add_edge(x: int, y: int) -> None:
    global gr
    gr[x].append(y)
    gr[y].append(x)
 
# Function to find the length of
# the shortest cycle in the graph
def shortest_cycle(n: int) -> int:
 
    # To store length of the shortest cycle
    ans = INT_MAX
 
    # For all vertices
    for i in range(n):
 
        # Make distance maximum
        dist = [int(1e9)] * n
 
        # Take a imaginary parent
        par = [-1] * n
 
        # Distance of source to source is 0
        dist[i] = 0
        q = deque()
 
        # Push the source element
        q.append(i)
 
        # Continue until queue is not empty
        while q:
 
            # Take the first element
            x = q[0]
            q.popleft()
 
            # Traverse for all it's childs
            for child in gr[x]:
 
                # If it is not visited yet
                if dist[child] == int(1e9):
 
                    # Increase distance by 1
                    dist[child] = 1 + dist[x]
 
                    # Change parent
                    par[child] = x
 
                    # Push into the queue
                    q.append(child)
 
                # If it is already visited
                elif par[x] != child and par[child] != x:
                    ans = min(ans, dist[x] +
                                   dist[child] + 1)
 
    # If graph contains no cycle
    if ans == INT_MAX:
        return -1
 
    # If graph contains cycle
    else:
        return ans
 
# Driver Code
if __name__ == "__main__":
 
    # Number of vertices
    n = 7
 
    # Add edges
    add_edge(0, 6)
    add_edge(0, 5)
    add_edge(5, 1)
    add_edge(1, 6)
    add_edge(2, 6)
    add_edge(2, 3)
    add_edge(3, 4)
    add_edge(4, 1)
 
    # Function call
    print(shortest_cycle(n))
 
# This code is contributed by
# sanjeev2552


C#




// C# implementation of the approach
using System;
using System.Collections.Generic;
 
class GFG
{
 
    static readonly int N = 100200;
    static List<int>[] gr = new List<int>[N];
 
    // Function to add edge
    static void Add_edge(int x, int y)
    {
        gr[x].Add(y);
        gr[y].Add(x);
    }
 
    // Function to find the length of
    // the shortest cycle in the graph
    static int shortest_cycle(int n)
    {
 
        // To store length of the shortest cycle
        int ans = int.MaxValue;
 
        // For all vertices
        for (int i = 0; i < n; i++)
        {
 
            // Make distance maximum
            int[] dist = new int[n];
            fill(dist, (int) 1e9);
 
            // Take a imaginary parent
            int[] par = new int[n];
            fill(par, -1);
 
            // Distance of source to source is 0
            dist[i] = 0;
            List<int> q = new List<int>();
 
            // Push the source element
            q.Add(i);
 
            // Continue until queue is not empty
            while (q.Count!=0)
            {
 
                // Take the first element
                int x = q[0];
                q.RemoveAt(0);
 
                // Traverse for all it's childs
                foreach (int child in gr[x])
                {
                    // If it is not visited yet
                    if (dist[child] == (int) (1e9))
                    {
 
                        // Increase distance by 1
                        dist[child] = 1 + dist[x];
 
                        // Change parent
                        par[child] = x;
 
                        // Push into the queue
                        q.Add(child);
                    } else if (par[x] != child && par[child] != x)
                        ans = Math.Min(ans, dist[x] + dist[child] + 1);
                }
            }
        }
 
        // If graph contains no cycle
        if (ans == int.MaxValue)
            return -1;
 
        // If graph contains cycle
        else
            return ans;
    }
    static int[] fill(int []arr, int val)
    {
        for(int i = 0;i<arr.GetLength(0);i++)
            arr[i] = val;
        return arr;
    }
                       
    // Driver Code
    public static void Main(String[] args)
    {
 
        for (int i = 0; i < N; i++)
            gr[i] = new List<int>();
 
        // Number of vertices
        int n = 7;
 
        // Add edges
        Add_edge(0, 6);
        Add_edge(0, 5);
        Add_edge(5, 1);
        Add_edge(1, 6);
        Add_edge(2, 6);
        Add_edge(2, 3);
        Add_edge(3, 4);
        Add_edge(4, 1);
 
        // Function call
        Console.WriteLine(shortest_cycle(n));
    }
}
 
// This code is contributed by 29AjayKumar


Javascript




<script>
 
// JavaScript implementation of the approach
var N = 100200;
var gr = Array.from(Array(N),()=>Array());
// Function to add edge
function Add_edge(x, y)
{
    gr[x].push(y);
    gr[y].push(x);
}
// Function to find the length of
// the shortest cycle in the graph
function shortest_cycle(n)
{
    // To store length of the shortest cycle
    var ans = 1000000000;
    // For all vertices
    for(var i = 0; i < n; i++)
    {
        // Make distance maximum
        var dist = Array(n).fill(1000000000);
        // Take a imaginary parent
        var par = Array(n).fill(-1);
     
        // Distance of source to source is 0
        dist[i] = 0;
        var q = [];
        // Push the source element
        q.push(i);
        // Continue until queue is not empty
        while (q.length!=0)
        {
            // Take the first element
            var x = q[0];
            q.shift();
            // Traverse for all it's childs
            for(var child of gr[x])
            {
                // If it is not visited yet
                if (dist[child] == 1000000000)
                {
                    // Increase distance by 1
                    dist[child] = 1 + dist[x];
                    // Change parent
                    par[child] = x;
                    // Push into the queue
                    q.push(child);
                } else if (par[x] != child && par[child] != x)
                    ans = Math.min(ans, dist[x] + dist[child] + 1);
            }
        }
    }
    // If graph contains no cycle
    if (ans == 1000000000)
        return -1;
    // If graph contains cycle
    else
        return ans;
}
function fill(arr, val)
{
    for(var i = 0;i<arr.length;i++)
        arr[i] = val;
    return arr;
}
                   
// Driver Code
// Number of vertices
var n = 7;
// push edges
Add_edge(0, 6);
Add_edge(0, 5);
Add_edge(5, 1);
Add_edge(1, 6);
Add_edge(2, 6);
Add_edge(2, 3);
Add_edge(3, 4);
Add_edge(4, 1);
// Function call
document.write(shortest_cycle(n));
 
 
</script>


Output: 

4

 

Time Complexity: O( |V| * (|V|+|E|)) for a graph G=(V, E) 

Memory Complexity: O(V^2) for a graph G=(V, E) 



Last Updated : 18 Mar, 2022
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads