Bellman-Ford vs Floyd-Warshall's algorithm: A Comparative Analysis

Last Updated : 05 Jan, 2024

Bellman-Ford Algorithm:

The Bellman-Ford algorithm is a single-source shortest-path algorithm that works by iteratively relaxing edges in the graph until the shortest path to all vertices is found. It is especially useful for graphs with negative edge weights, as it can detect negative cycles and return a suitable error message.

Floyd-Warshall Algorithm:

The Floyd-Warshall algorithm is a multi-source shortest path algorithm that works by computing the shortest path between all pairs of vertices in the graph using dynamic programming. It is known for its simplicity and ease of implementation, making it a popular choice in many applications.

In this article, we will compare and contrast the Bellman-Ford algorithm and the Floyd-Warshall algorithm, examining their respective strengths and weaknesses, time and space complexity, and implementation details.

As we know that both of these algorithms work on Directed Graphs, Let us take a common example to understand the approaches of these algorithms.

Bellman-ford's Algorithm

The Bellman-Ford algorithm can be implemented as follows:

Initialize the distance from the source vertex (i.e. 0) to all other vertices in the graph to infinity, except for the distance from the source vertex to itself, which is 0.
For each vertex in the graph, repeat the following process |V|-1 times (where |V| is the number of vertices in the graph):
For each edge (u, v) in the graph, where u is the source vertex and v is the destination vertex, relax the edge by updating the distance to v if the distance to u plus the weight of (u, v) is less than the current distance to v.
After the (|V|-1)th iteration, check for the presence of negative cycles by iterating over all the edges in the graph and checking if any of them can still be relaxed. If so, then a negative cycle is present, and the algorithm returns a suitable error message.

Here's how the algorithm would work:

Initialize the distance from A to all other vertices to infinity, except for the distance from A to itself, which is 0: dist[0] = 0.
Repeat the process of relaxing edges 5 times (|V|-1 = 6-1 = 5) to relax all edges in the graph.
Check for negative cycles by iterating over all edges in the graph. In this case, there are no negative cycles, as all edges have been relaxed in the previous step.

Hence our distance from source 0 to each node will look like this,

Implementation of Bellman-Ford's Algorithm:

C++

// C++ code for the above approach:
#include <bits/stdc++.h>

using namespace std;
class Solution {
public:
    void shortest_paths(vector<vector<int> >& grid, int V)
    {
        vector<int> dist(V, 1e9);

        // Taking 0 as our source
        dist[0] = 0;

        // Relaxing V-1 times
        for (int i = 0; i < V; i++) {

            // For each value in the given
            // vector of [u, v, wt]
            for (auto it : grid) {
                int u = it[0];
                int v = it[1];
                int wt = it[2];

                // If there exist a better
                // distance and pvs dist
                // of u should not
                // be infinite
                if (dist[u] != 1e9
                    && dist[u] + wt < dist[v]) {
                    dist[v] = dist[u] + wt;
                }
            }
        }

        // If negetive cycle exist then
        // it will still reduce for
        // Vth traversal
        for (auto it : grid) {
            int u = it[0];
            int v = it[1];
            int wt = it[2];
            if (dist[u] != 1e9 && dist[u] + wt < dist[v]) {
                cout << "ERROR ALERT, Negetive Cycle exists"
                     << endl;
            }
        }
        cout << "The shortest distances from " << endl;
        for (int i = 0; i < V; i++) {
            cout << " 0 to " << i << " ---> " << dist[i]
                 << endl;
        }
    }
};

// Driver code
int main()
{
    int V = 6;
    vector<vector<int> > grid
        = { { 0, 1, 2 }, { 0, 3, 5 }, { 0, 4, 3 }, { 1, 0, 3 }, { 1, 5, 6 }, { 1, 2, 2 }, { 1, 3, 2 }, { 2, 5, 1 }, { 2, 3, 1 }, { 3, 4, 1 }, { 4, 3, 2 } };
    Solution s1;

    // Function call
    s1.shortest_paths(grid, V);
    return 0;
}

Java

import java.util.Arrays;
import java.util.List;
import java.util.ArrayList;

class Solution {
    public void shortestPaths(List<List<Integer>> grid, int V) {
        int[] dist = new int[V];
        Arrays.fill(dist, Integer.MAX_VALUE);

        // Taking 0 as our source
        dist[0] = 0;

        // Relaxing V-1 times
        for (int i = 0; i < V; i++) {

            // For each value in the given
            // list of [u, v, wt]
            for (List<Integer> it : grid) {
                int u = it.get(0);
                int v = it.get(1);
                int wt = it.get(2);

                // If there exists a better
                // distance and the previous distance
                // of u should not be infinite
                if (dist[u] != Integer.MAX_VALUE
                        && dist[u] + wt < dist[v]) {
                    dist[v] = dist[u] + wt;
                }
            }
        }

        // If a negative cycle exists, it will still reduce for
        // Vth traversal
        for (List<Integer> it : grid) {
            int u = it.get(0);
            int v = it.get(1);
            int wt = it.get(2);
            if (dist[u] != Integer.MAX_VALUE && dist[u] + wt < dist[v]) {
                System.out.println("ERROR ALERT, Negative Cycle exists");
            }
        }
        System.out.println("The shortest distances from:");
        for (int i = 0; i < V; i++) {
            System.out.println("0 to " + i + " ---> " + dist[i]);
        }
    }
}

public class Main {
    public static void main(String[] args) {
        int V = 6;
        List<List<Integer>> grid = new ArrayList<>();
        grid.add(Arrays.asList(0, 1, 2));
        grid.add(Arrays.asList(0, 3, 5));
        grid.add(Arrays.asList(0, 4, 3));
        grid.add(Arrays.asList(1, 0, 3));
        grid.add(Arrays.asList(1, 5, 6));
        grid.add(Arrays.asList(1, 2, 2));
        grid.add(Arrays.asList(1, 3, 2));
        grid.add(Arrays.asList(2, 5, 1));
        grid.add(Arrays.asList(2, 3, 1));
        grid.add(Arrays.asList(3, 4, 1));
        grid.add(Arrays.asList(4, 3, 2));

        Solution s1 = new Solution();

        // Function call
        s1.shortestPaths(grid, V);
    }
}

Python

# python code for the above approach:
class Solution:
    def shortest_paths(self, grid, V):
        dist = [float('inf')] * V

        # Taking 0 as our source
        dist[0] = 0

        # Relaxing V-1 times
        for i in range(V):
            # For each value in the given
            # list of [u, v, wt]
            for u, v, wt in grid:
                # If there exists a better
                # distance and previous dist
                # of u should not be infinite
                if dist[u] != float('inf') and dist[u] + wt < dist[v]:
                    dist[v] = dist[u] + wt

        # If negative cycle exists then
        # it will still reduce for
        # Vth traversal
        for u, v, wt in grid:
            if dist[u] != float('inf') and dist[u] + wt < dist[v]:
                print("ERROR ALERT, Negative Cycle exists")

        print("The shortest distances from ")
        for i in range(V):
            print(" 0 to", i, "--->", dist[i])


# Driver code
if __name__ == "__main__":
    V = 6
    grid = [[0, 1, 2], [0, 3, 5], [0, 4, 3], [1, 0, 3], [1, 5, 6], [
        1, 2, 2], [1, 3, 2], [2, 5, 1], [2, 3, 1], [3, 4, 1], [4, 3, 2]]
    s1 = Solution()

    # Function call
    s1.shortest_paths(grid, V)

using System;
using System.Collections.Generic;

class Solution
{
    public void ShortestPaths(List<List<int>> grid, int V)
    {
        List<int> dist = new List<int>(new int[V]);

        // Initialize distances to maximum value
        for (int i = 0; i < V; i++)
        {
            dist[i] = int.MaxValue;
        }

        // Taking 0 as our source
        dist[0] = 0;

        // Relaxing V-1 times
        for (int i = 0; i < V - 1; i++)
        {
            // For each value in the given
            // list of [u, v, wt]
            foreach (var it in grid)
            {
                int u = it[0];
                int v = it[1];
                int wt = it[2];

                // If there exists a better
                // distance and the previous dist
                // of u should not be infinite
                if (dist[u] != int.MaxValue
                    && dist[u] + wt < dist[v])
                {
                    dist[v] = dist[u] + wt;
                }
            }
        }

        // Check for negative cycles
        for (int i = 0; i < V; i++)
        {
            foreach (var it in grid)
            {
                int u = it[0];
                int v = it[1];
                int wt = it[2];

                if (dist[u] != int.MaxValue && dist[u] + wt < dist[v])
                {
                    Console.WriteLine("ERROR ALERT, Negative Cycle exists");
                    return;
                }
            }
        }

        Console.WriteLine("The shortest distances from ");
        for (int i = 0; i < V; i++)
        {
            Console.WriteLine($"0 to {i} ---> {(dist[i] == int.MaxValue ? "INF" : dist[i].ToString())}");
        }
    }
}

class Program
{
    static void Main()
    {
        int V = 6;
        List<List<int>> grid = new List<List<int>>
        {
            new List<int> { 0, 1, 2 },
            new List<int> { 0, 3, 5 },
            new List<int> { 0, 4, 3 },
            new List<int> { 1, 0, 3 },
            new List<int> { 1, 5, 6 },
            new List<int> { 1, 2, 2 },
            new List<int> { 1, 3, 2 },
            new List<int> { 2, 5, 1 },
            new List<int> { 2, 3, 1 },
            new List<int> { 3, 4, 1 },
            new List<int> { 4, 3, 2 }
        };

        Solution s1 = new Solution();

        // Function call
        s1.ShortestPaths(grid, V);
    }
}

JavaScript

class Solution {
    shortestPaths(grid, V) {
        const dist = new Array(V).fill(1e9);

        // Taking 0 as our source
        dist[0] = 0;

        // Relaxing V-1 times
        for (let i = 0; i < V; i++) {

            // For each value in the given vector of [u, v, wt]
            for (const it of grid) {
                const u = it[0];
                const v = it[1];
                const wt = it[2];

                // If there exists a better distance and the previous distance
                // of u should not be infinite
                if (dist[u] !== 1e9 && dist[u] + wt < dist[v]) {
                    dist[v] = dist[u] + wt;
                }
            }
        }

        // If a negative cycle exists, it will still reduce for Vth traversal
        for (const it of grid) {
            const u = it[0];
            const v = it[1];
            const wt = it[2];
            if (dist[u] !== 1e9 && dist[u] + wt < dist[v]) {
                console.log("ERROR ALERT, Negative Cycle exists");
            }
        }
        console.log("The shortest distances from:");
        for (let i = 0; i < V; i++) {
            console.log(`0 to ${i} ---> ${dist[i]}`);
        }
    }
}

// Driver code
const V = 6;
const grid = [
    [0, 1, 2],
    [0, 3, 5],
    [0, 4, 3],
    [1, 0, 3],
    [1, 5, 6],
    [1, 2, 2],
    [1, 3, 2],
    [2, 5, 1],
    [2, 3, 1],
    [3, 4, 1],
    [4, 3, 2]
];
const s1 = new Solution();

// Function call
s1.shortestPaths(grid, V);

Output

The shortest distances from 
 0 to 0 ---> 0
 0 to 1 ---> 2
 0 to 2 ---> 4
 0 to 3 ---> 4
 0 to 4 ---> 3
 0 to 5 ---> 5

Time complexity: O(|V||E|)
Auxiliary space: O(|V|)

Floyd-Warshall's Algorithm

The Floyd-Warshall algorithm is used to find the shortest path between all pairs of nodes in a weighted graph. It works by maintaining a distance matrix where each entry (i, j) represents the shortest distance from node i to node j. The algorithm starts by initializing the distance matrix with the weights of the edges in the graph. Then, it iteratively updates the distance matrix by considering all possible intermediate nodes and choosing the path with the minimum total weight.

Implementation of Floyd-Warshall's Algorithm:

C++

// C++ code for the above approach:
#include <bits/stdc++.h>
using namespace std;
class Solution {
public:
    void shortest_dist(vector<vector<int> >& matrix)
    {
        int N = matrix.size();

        // changing value of -1
        // to infinite
        for (int i = 0; i < N; i++) {
            for (int j = 0; j < N; j++) {
                if (matrix[i][j] == -1) {
                    matrix[i][j] = 1e9;
                }

                // The distance of node to
                // itself will be 0.
                if (i == j) {
                    matrix[i][j] = 0;
                }
            }
        }

        // Assign the better distance
        // (if exist) in the matrix
        // one by one making each node
        // as our intermediate node
        for (int via = 0; via < N; via++) {
            for (int i = 0; i < N; i++) {
                for (int j = 0; j < N; j++) {
                    matrix[i][j] = min(
                        matrix[i][j],
                        matrix[i][via] + matrix[via][j]);
                }
            }
        }

        // Replacing infinity to -1,
        // after the work is done
        for (int i = 0; i < N; i++) {
            for (int j = 0; j < N; j++) {
                if (matrix[i][j] == 1e9) {
                    matrix[i][j] = -1;
                }
            }
        }
    }
};

// Driver code
int main()
{
    vector<vector<int> > grid
        = { { 0, 1, 2 }, { 0, 3, 5 }, { 0, 4, 3 }, { 1, 0, 3 }, { 1, 5, 6 }, { 1, 2, 2 }, { 1, 3, 2 }, { 2, 5, 1 }, { 2, 3, 1 }, { 3, 4, 1 }, { 4, 3, 2 } };
    int V = 6;
    vector<vector<int> > matrix(V, vector<int>(V, -1));

    // creating matrix representation
    for (auto it : grid) {
        int u = it[0];
        int v = it[1];
        int wt = it[2];
        matrix[u][v] = wt;
    }

    Solution obj;

    // Function call
    obj.shortest_dist(matrix);

    for (auto row : matrix) {
        for (auto it : row) {
            if (it == 1e9) {
                it = -1;
            }
            cout << it << "\t";
        }
        cout << endl;
    }

    return 0;
}

Java

import java.util.Arrays;

public class Solution {
    public void shortestDist(int[][] matrix) {
        int N = matrix.length;

        // Changing value of -1 to infinite
        for (int i = 0; i < N; i++) {
            for (int j = 0; j < N; j++) {
                if (matrix[i][j] == -1) {
                    matrix[i][j] = Integer.MAX_VALUE;
                }

                // The distance of node to itself will be 0
                if (i == j) {
                    matrix[i][j] = 0;
                }
            }
        }

        // Assign the better distance (if exist) in the matrix
        // one by one making each node as our intermediate node
        for (int via = 0; via < N; via++) {
            for (int i = 0; i < N; i++) {
                for (int j = 0; j < N; j++) {
                    if (matrix[i][via] != Integer.MAX_VALUE && matrix[via][j] != Integer.MAX_VALUE &&
                            matrix[i][j] > matrix[i][via] + matrix[via][j]) {
                        matrix[i][j] = matrix[i][via] + matrix[via][j];
                    }
                }
            }
        }

        // Replacing infinity to -1, after the work is done
        for (int i = 0; i < N; i++) {
            for (int j = 0; j < N; j++) {
                if (matrix[i][j] == Integer.MAX_VALUE) {
                    matrix[i][j] = -1;
                }
            }
        }
    }

    // Driver code
    public static void main(String[] args) {
        int[][] grid = {
            {0, 1, 2}, {0, 3, 5}, {0, 4, 3}, {1, 0, 3}, {1, 5, 6},
            {1, 2, 2}, {1, 3, 2}, {2, 5, 1}, {2, 3, 1}, {3, 4, 1}, {4, 3, 2}
        };
        int V = 6;
        int[][] matrix = new int[V][V];

        // Initializing matrix representation
        for (int[] it : matrix) {
            Arrays.fill(it, -1);
        }

        // Creating matrix representation
        for (int[] it : grid) {
            int u = it[0];
            int v = it[1];
            int wt = it[2];
            matrix[u][v] = wt;
        }

        Solution obj = new Solution();

        // Function call
        obj.shortestDist(matrix);

        // Printing the result
        for (int[] row : matrix) {
            for (int it : row) {
                if (it == -1) {
                    System.out.print(it + "\t");
                } else {
                    System.out.print(it + "\t");
                }
            }
            System.out.println();
        }
    }
}

Python3

# Python code for the above approach:

class Solution:
    def shortest_dist(self, matrix):
        N = len(matrix)

        # changing value of -1 
        # to infinite
        for i in range(N):
            for j in range(N):
                if matrix[i][j] == -1:
                    matrix[i][j] = float('inf')

                # The distance of node to 
                # itself will be 0.
                if i == j:
                    matrix[i][j] = 0

        # Assign the better distance 
        # (if exist) in the matrix
        # one by one making each node 
        # as our intermediate node
        for via in range(N):
            for i in range(N):
                for j in range(N):
                    matrix[i][j] = min(matrix[i][j], matrix[i][via] + matrix[via][j])

        # Replacing infinity to -1, 
        # after the work is done
        for i in range(N):
            for j in range(N):
                if matrix[i][j] == float('inf'):
                    matrix[i][j] = -1

# Driver code
if __name__ == "__main__":
    grid = [[0, 1, 2], [0, 3, 5], [0, 4, 3], [1, 0, 3], [1, 5, 6], [1, 2, 2], [1, 3, 2], [2, 5, 1], [2, 3, 1], [3, 4, 1], [4, 3, 2]]
    V = 6
    matrix = [[-1 for _ in range(V)] for _ in range(V)]

    # Creating matrix representation
    for it in grid:
        u, v, wt = it[0], it[1], it[2]
        matrix[u][v] = wt

    obj = Solution()

    # Function call
    obj.shortest_dist(matrix)

    for row in matrix:
        for it in row:
            if it == -1:
                it = -1
            print(it, end="\t")
        print()


# This code is contributed by Pushpesh raj

using System;
using System.Collections.Generic;

public class Solution
{
    public void ShortestDist(List<List<int>> matrix)
    {
        int N = matrix.Count;

        // Convert -1 to infinite and set distance to self as 0
        for (int i = 0; i < N; i++)
        {
            for (int j = 0; j < N; j++)
            {
                if (matrix[i][j] == -1)
                    matrix[i][j] = int.MaxValue;
                
                if (i == j)
                    matrix[i][j] = 0;
            }
        }

        // Applying Floyd-Warshall algorithm
        for (int k = 0; k < N; k++)
        {
            for (int i = 0; i < N; i++)
            {
                for (int j = 0; j < N; j++)
                {
                    if (matrix[i][k] != int.MaxValue && matrix[k][j] != int.MaxValue && 
                        matrix[i][k] + matrix[k][j] < matrix[i][j])
                    {
                        matrix[i][j] = matrix[i][k] + matrix[k][j];
                    }
                }
            }
        }

        // Convert back to -1 from infinite
        for (int i = 0; i < N; i++)
        {
            for (int j = 0; j < N; j++)
            {
                if (matrix[i][j] == int.MaxValue)
                    matrix[i][j] = -1;
            }
        }
    }
}

public class Program
{
    public static void Main()
    {
        List<List<int>> grid = new List<List<int>>
        {
            new List<int> { 0, 1, 2 }, new List<int> { 0, 3, 5 }, new List<int> { 0, 4, 3 },
            new List<int> { 1, 0, 3 }, new List<int> { 1, 5, 6 }, new List<int> { 1, 2, 2 },
            new List<int> { 1, 3, 2 }, new List<int> { 2, 5, 1 }, new List<int> { 2, 3, 1 },
            new List<int> { 3, 4, 1 }, new List<int> { 4, 3, 2 }
        };

        int V = 6;
        List<List<int>> matrix = new List<List<int>>(V);
        for (int i = 0; i < V; i++)
        {
            matrix.Add(new List<int>(new int[V]));
            for (int j = 0; j < V; j++)
            {
                matrix[i][j] = -1;
            }
        }

        // Create matrix representation
        foreach (var it in grid)
        {
            int u = it[0];
            int v = it[1];
            int wt = it[2];
            matrix[u][v] = wt;
        }

        Solution obj = new Solution();

        // Call the function
        obj.ShortestDist(matrix);

        // Print the matrix
        foreach (var row in matrix)
        {
            foreach (var value in row)
            {
                Console.Write($"{value}\t");
            }
            Console.WriteLine();
        }
    }
}


// This code is contributed by shivamgupta310570

JavaScript

class Solution {
    shortestDist(matrix) {
        const N = matrix.length;

        // Changing value of -1 to infinite
        for (let i = 0; i < N; i++) {
            for (let j = 0; j < N; j++) {
                if (matrix[i][j] === -1) {
                    matrix[i][j] = Infinity;
                }

                // The distance of node to itself will be 0.
                if (i === j) {
                    matrix[i][j] = 0;
                }
            }
        }

        // Assign the better distance (if exist) in the matrix
        // one by one making each node as our intermediate node
        for (let via = 0; via < N; via++) {
            for (let i = 0; i < N; i++) {
                for (let j = 0; j < N; j++) {
                    matrix[i][j] = Math.min(
                        matrix[i][j],
                        matrix[i][via] + matrix[via][j]
                    );
                }
            }
        }

        // Replacing infinity with -1 after the work is done
        for (let i = 0; i < N; i++) {
            for (let j = 0; j < N; j++) {
                if (matrix[i][j] === Infinity) {
                    matrix[i][j] = -1;
                }
            }
        }
    }
}

// Driver code
const grid = [
    [0, 1, 2], [0, 3, 5], [0, 4, 3], [1, 0, 3], [1, 5, 6],
    [1, 2, 2], [1, 3, 2], [2, 5, 1], [2, 3, 1], [3, 4, 1], [4, 3, 2]
];

const V = 6;
const matrix = Array.from({ length: V }, () => Array(V).fill(-1));

// Creating matrix representation
for (const it of grid) {
    const [u, v, wt] = it;
    matrix[u][v] = wt;
}

const obj = new Solution();

// Function call
obj.shortestDist(matrix);

for (const row of matrix) {
    for (const it of row) {
        if (it === -1) {
            console.log("-1\t");
        } else {
            console.log(it + "\t");
        }
    }
    console.log();
}

Output

0    2    4    4    3    5    
3    0    2    2    3    3    
-1    -1    0    1    2    1    
-1    -1    -1    0    1    -1    
-1    -1    -1    2    0    -1    
-1    -1    -1    -1    -1    0

Time complexity: O(N^3)
Auxiliary Space: O(N^2)

Analysis of Algorithms

risheekshukla10

Improve

Article Tags :

Practice Tags :

Bellman-Ford vs Floyd-Warshall's algorithm: A Comparative Analysis

Bellman-Ford Algorithm:

Floyd-Warshall Algorithm:

Bellman-ford's Algorithm

Implementation of Bellman-Ford's Algorithm:

Floyd-Warshall's Algorithm

Implementation of Floyd-Warshall's Algorithm:

Similar Reads

Basics & Prerequisites

Data Structures

Algorithms

Advanced

Interview Preparation

Practice Problem

Thank You!

What kind of Experience do you want to share?