Boruvka’s algorithm | Greedy Algo-9

Last Updated : 17 Mar, 2023

We have discussed the following topics on Minimum Spanning Tree.
Applications of Minimum Spanning Tree Problem
Kruskal’s Minimum Spanning Tree Algorithm
Prim’s Minimum Spanning Tree Algorithm
In this post, Boruvka’s algorithm is discussed. Like Prim’s and Kruskal’s, Boruvka’s algorithm is also a Greedy algorithm. Below is a complete algorithm.

1) Input is a connected, weighted and un-directed graph.
2) Initialize all vertices as individual components (or sets).
3) Initialize MST as empty.
4) While there are more than one components, do following
   for each component.
      a)  Find the closest weight edge that connects this 
          component to any other component.
      b)  Add this closest edge to MST if not already added.  
5) Return MST.

Below is the idea behind the above algorithm (The idea is the same as Prim’s MST algorithm).
A spanning tree means all vertices must be connected. So the two disjoint subsets (discussed above) of vertices must be connected to make a Spanning Tree. And they must be connected with the minimum weight edge to make it a Minimum Spanning Tree.
Let us understand the algorithm in the below example.

Boruvka's algorithm Image 1

Initially, MST is empty. Every vertex is single component as highlighted in blue color in the below diagram.

Boruvka's algorithm Image 2

For every component, find the cheapest edge that connects it to some other component.

Component                Cheapest Edge that connects 
                         it to some other component
  {0}                           0-1
  {1}                           0-1
  {2}                           2-8
  {3}                           2-3
  {4}                           3-4
  {5}                           5-6
  {6}                           6-7
  {7}                           6-7
  {8}                           2-8

The cheapest edges are highlighted with green color. Now MST becomes {0-1, 2-8, 2-3, 3-4, 5-6, 6-7}.

Boruvka's algorithm Image 3

After above step, components are {{0,1}, {2,3,4,8}, {5,6,7}}. The components are encircled with blue color.

Boruvka's algorithm Image 4

We again repeat the step, i.e., for every component, find the cheapest edge that connects it to some other component.

Component                Cheapest Edge that connects 
                         it to some other component
  {0,1}                        1-2 (or 0-7)
  {2,3,4,8}                    2-5
  {5,6,7}                      2-5

The cheapest edges are highlighted with green color. Now MST becomes {0-1, 2-8, 2-3, 3-4, 5-6, 6-7, 1-2, 2-5}

Boruvka's algorithm Image 5

At this stage, there is only one component {0, 1, 2, 3, 4, 5, 6, 7, 8} which has all edges. Since there is only one component left, we stop and return MST.

Implementation: Below is the implementation of the above algorithm. The input graph is represented as a collection of edges and union-find data structure is used to keep track of components.

C++

// Boruvka's algorithm to find Minimum Spanning
// Tree of a given connected, undirected and weighted graph
#include <bits/stdc++.h>
using namespace std;
 
// Class to represent a graph
class Graph {
    int V; // No. of vertices
    vector<vector<int> >graph; // default dictionary to store graph
 
    // A utility function to find set of an element i
    // (uses path compression technique)
    int find(vector<int>& parent, int i)
    {
        if (parent[i] == i) {
            return i;
        }
        return find(parent, parent[i]);
    }
 
    // A function that does union of two sets of x and y
    // (uses union by rank)
    void unionSet(vector<int>& parent, vector<int>& rank,
                  int x, int y)
    {
        int xroot = find(parent, x);
        int yroot = find(parent, y);
 
        // Attach smaller rank tree under root of high rank
        // tree (Union by Rank)
        if (rank[xroot] < rank[yroot]) {
            parent[xroot] = yroot;
        }
        else if (rank[xroot] > rank[yroot]) {
            parent[yroot] = xroot;
        }
        // If ranks are same, then make one as root and
        // increment its rank by one
        else {
            parent[yroot] = xroot;
            rank[xroot]++;
        }
    }
 
public:
    Graph(int vertices)
    {
        V = vertices;
        graph = vector<vector<int> >();
    }
 
    // function to add an edge to graph
    void addEdge(int u, int v, int w)
    {
        graph.push_back({ u, v, w });
    }
 
    // The main function to construct MST using Kruskal's
    // algorithm
    void boruvkaMST()
    {
        vector<int> parent(V);
 
        // An array to store index of the cheapest edge of
        // subset. It store [u,v,w] for each component
        vector<int> rank(V);
        vector<vector<int> > cheapest(V,
                                      vector<int>(3, -1));
 
        // Initially there are V different trees.
        // Finally there will be one tree that will be MST
        int numTrees = V;
        int MSTweight = 0;
 
        // Create V subsets with single elements
        for (int node = 0; node < V; node++) {
            parent[node] = node;
            rank[node] = 0;
        }
 
        // Keep combining components (or sets) until all
        // components are not combined into single MST
        while (numTrees > 1) {
 
            // Traverse through all edges and update
            // cheapest of every component
            for (int i = 0; i < graph.size(); i++) {
 
                // Find components (or sets) of two corners
                // of current edge
                int u = graph[i][0], v = graph[i][1],
                    w = graph[i][2];
                int set1 = find(parent, u),
                    set2 = find(parent, v);
 
                // If two corners of current edge belong to
                // same set, ignore current edge. Else check
                // if current edge is closer to previous
                // cheapest edges of set1 and set2
                if (set1 != set2) {
                    if (cheapest[set1][2] == -1
                        || cheapest[set1][2] > w) {
                        cheapest[set1] = { u, v, w };
                    }
                    if (cheapest[set2][2] == -1
                        || cheapest[set2][2] > w) {
                        cheapest[set2] = { u, v, w };
                    }
                }
            }
 
            // Consider the above picked cheapest edges and
            // add them to MST
            for (int node = 0; node < V; node++) {
 
                // Check if cheapest for current set exists
                if (cheapest[node][2] != -1) {
                    int u = cheapest[node][0],
                        v = cheapest[node][1],
                        w = cheapest[node][2];
                    int set1 = find(parent, u),
                        set2 = find(parent, v);
                    if (set1 != set2) {
                        MSTweight += w;
                        unionSet(parent, rank, set1, set2);
                        printf("Edge %d-%d with weight %d "
                               "included in MST\n",
                               u, v, w);
                        numTrees--;
                    }
                }
            }
            for (int node = 0; node < V; node++) {
 
                // reset cheapest array
                cheapest[node][2] = -1;
            }
        }
        printf("Weight of MST is %d\n", MSTweight);
    }
};
int main()
{
    Graph g(4);
    g.addEdge(0, 1, 10);
    g.addEdge(0, 2, 6);
    g.addEdge(0, 3, 5);
    g.addEdge(1, 3, 15);
    g.addEdge(2, 3, 4);
 
    g.boruvkaMST();
}
// This code is contributed by prajwal kandekar

Java

// Boruvka's algorithm to find Minimum Spanning
// Tree of a given connected, undirected and weighted graph
import java.util.*;
 
// Class to represent a graph
class Graph {
  private int V; // No. of vertices
  private List<List<Integer> >
    graph; // default dictionary to store graph
 
  Graph(int vertices)
  {
    V = vertices;
    graph = new ArrayList<>();
  }
 
  // function to add an edge to graph
  void addEdge(int u, int v, int w)
  {
    graph.add(Arrays.asList(u, v, w));
  }
 
  // A utility function to find set of an element i
  // (uses path compression technique)
  private int find(List<Integer> parent, int i)
  {
    if (parent.get(i) == i) {
      return i;
    }
    return find(parent, parent.get(i));
  }
 
  // A function that does union of two sets of x and y
  // (uses union by rank)
  private void unionSet(List<Integer> parent,
                        List<Integer> rank, int x, int y)
  {
    int xroot = find(parent, x);
    int yroot = find(parent, y);
 
    // Attach smaller rank tree under root of high rank
    // tree (Union by Rank)
    if (rank.get(xroot) < rank.get(yroot)) {
      parent.set(xroot, yroot);
    }
    else if (rank.get(xroot) > rank.get(yroot)) {
      parent.set(yroot, xroot);
    }
 
    // If ranks are same, then make one as root and
    // increment its rank by one
    else {
      parent.set(yroot, xroot);
      rank.set(xroot, rank.get(xroot) + 1);
    }
  }
 
  // The main function to construct MST using Kruskal's
  // algorithm
  void boruvkaMST()
  {
    List<Integer> parent = new ArrayList<>();
 
    // An array to store index of the cheapest edge of
    // subset. It store [u,v,w] for each component
    List<Integer> rank = new ArrayList<>();
 
    List<List<Integer> > cheapest = new ArrayList<>();
 
    // Initially there are V different trees.
    // Finally there will be one tree that will be MST
    int numTrees = V;
    int MSTweight = 0;
 
    // Create V subsets with single elements
    for (int node = 0; node < V; node++) {
      parent.add(node);
      rank.add(0);
      cheapest.add(Arrays.asList(-1, -1, -1));
    }
 
    // Keep combining components (or sets) until all
    // components are not combined into single MST
    while (numTrees > 1) {
 
      // Traverse through all edges and update
      // cheapest of every component
      for (List<Integer> edge : graph) {
 
        // Find components (or sets) of two corners
        // of current edge
        int u = edge.get(0), v = edge.get(1),
        w = edge.get(2);
        int set1 = find(parent, u),
        set2 = find(parent, v);
 
        // If two corners of current edge belong to
        // same set, ignore current edge. Else check
        // if current edge is closer to previous
        // cheapest edges of set1 and set2
        if (set1 != set2) {
          if (cheapest.get(set1).get(2) == -1
              || cheapest.get(set1).get(2) > w) {
            cheapest.set(
              set1, Arrays.asList(u, v, w));
          }
          if (cheapest.get(set2).get(2) == -1
              || cheapest.get(set2).get(2) > w) {
            cheapest.set(
              set2, Arrays.asList(u, v, w));
          }
        }
      }
 
      // Consider the above picked cheapest edges and
      // add them to MST
      for (int node = 0; node < V; node++) {
 
        // Check if cheapest for current set exists
        if (cheapest.get(node).get(2) != -1) {
          int u = cheapest.get(node).get(0),
          v = cheapest.get(node).get(1),
          w = cheapest.get(node).get(2);
          int set1 = find(parent, u),
          set2 = find(parent, v);
          if (set1 != set2) {
            MSTweight += w;
            unionSet(parent, rank, set1, set2);
            System.out.printf(
              "Edge %d-%d with weight %d included in MST\n",
              u, v, w);
            numTrees--;
          }
        }
      }
      for (List<Integer> list : cheapest) {
        // reset cheapest array
        list.set(2, -1);
      }
    }
    System.out.printf("Weight of MST is %d\n",
                      MSTweight);
  }
}
 
class GFG {
  public static void main(String[] args)
  {
    Graph g = new Graph(4);
    g.addEdge(0, 1, 10);
    g.addEdge(0, 2, 6);
    g.addEdge(0, 3, 5);
    g.addEdge(1, 3, 15);
    g.addEdge(2, 3, 4);
    g.boruvkaMST();
  }
}
// This code is contributed by prasad264

Python

# Boruvka's algorithm to find Minimum Spanning
# Tree of a given connected, undirected and weighted graph
 
from collections import defaultdict
 
#Class to represent a graph
class Graph:
 
    def __init__(self,vertices):
        self.V= vertices #No. of vertices
        self.graph = [] # default dictionary to store graph
         
  
    # function to add an edge to graph
    def addEdge(self,u,v,w):
        self.graph.append([u,v,w])
 
    # A utility function to find set of an element i
    # (uses path compression technique)
    def find(self, parent, i):
        if parent[i] == i:
            return i
        return self.find(parent, parent[i])
 
    # A function that does union of two sets of x and y
    # (uses union by rank)
    def union(self, parent, rank, x, y):
        xroot = self.find(parent, x)
        yroot = self.find(parent, y)
 
        # Attach smaller rank tree under root of high rank tree
        # (Union by Rank)
        if rank[xroot] < rank[yroot]:
            parent[xroot] = yroot
        elif rank[xroot] > rank[yroot]:
            parent[yroot] = xroot
        #If ranks are same, then make one as root and increment
        # its rank by one
        else :
            parent[yroot] = xroot
            rank[xroot] += 1
 
    # The main function to construct MST using Kruskal's algorithm
    def boruvkaMST(self):
        parent = []; rank = []; 
 
        # An array to store index of the cheapest edge of
        # subset. It store [u,v,w] for each component
        cheapest =[]
 
        # Initially there are V different trees.
        # Finally there will be one tree that will be MST
        numTrees = self.V
        MSTweight = 0
 
        # Create V subsets with single elements
        for node in range(self.V):
            parent.append(node)
            rank.append(0)
            cheapest =[-1] * self.V
     
        # Keep combining components (or sets) until all
        # components are not combined into single MST
 
        while numTrees > 1:
 
            # Traverse through all edges and update
               # cheapest of every component
            for i in range(len(self.graph)):
 
                # Find components (or sets) of two corners
                # of current edge
                u,v,w =  self.graph[i]
                set1 = self.find(parent, u)
                set2 = self.find(parent ,v)
 
                # If two corners of current edge belong to
                # same set, ignore current edge. Else check if 
                # current edge is closer to previous
                # cheapest edges of set1 and set2
                if set1 != set2:     
                     
                    if cheapest[set1] == -1 or cheapest[set1][2] > w :
                        cheapest[set1] = [u,v,w] 
 
                    if cheapest[set2] == -1 or cheapest[set2][2] > w :
                        cheapest[set2] = [u,v,w]
 
            # Consider the above picked cheapest edges and add them
            # to MST
            for node in range(self.V):
 
                #Check if cheapest for current set exists
                if cheapest[node] != -1:
                    u,v,w = cheapest[node]
                    set1 = self.find(parent, u)
                    set2 = self.find(parent ,v)
 
                    if set1 != set2 :
                        MSTweight += w
                        self.union(parent, rank, set1, set2)
                        print ("Edge %d-%d with weight %d included in MST" % (u,v,w))
                        numTrees = numTrees - 1
             
            #reset cheapest array
            cheapest =[-1] * self.V
 
             
        print ("Weight of MST is %d" % MSTweight)
                           
 
     
g = Graph(4)
g.addEdge(0, 1, 10)
g.addEdge(0, 2, 6)
g.addEdge(0, 3, 5)
g.addEdge(1, 3, 15)
g.addEdge(2, 3, 4)
 
g.boruvkaMST()
 
#This code is contributed by Neelam Yadav

C#

// Boruvka's algorithm to find Minimum Spanning
// Tree of a given connected, undirected and weighted graph
using System;
using System.Collections.Generic;
 
// Class to represent a graph
class Graph {
    private int V; // No. of vertices
    private List<List<int> >
        graph; // default dictionary to store graph
 
    // A utility function to find set of an element i
    // (uses path compression technique)
    private int Find(List<int> parent, int i)
    {
        if (parent[i] == i) {
            return i;
        }
        return Find(parent, parent[i]);
    }
 
    // A function that does union of two sets of x and y
    // (uses union by rank)
    private void UnionSet(List<int> parent, List<int> rank,
                          int x, int y)
    {
        int xroot = Find(parent, x);
        int yroot = Find(parent, y);
 
        // Attach smaller rank tree under root of high rank
        // tree (Union by Rank)
        if (rank[xroot] < rank[yroot]) {
            parent[xroot] = yroot;
        }
        else if (rank[xroot] > rank[yroot]) {
            parent[yroot] = xroot;
        }
 
        // If ranks are same, then make one as root and
        // increment its rank by one
        else {
            parent[yroot] = xroot;
            rank[xroot]++;
        }
    }
 
    public Graph(int vertices)
    {
        V = vertices;
        graph = new List<List<int> >();
    }
 
    // function to add an edge to graph
    public void AddEdge(int u, int v, int w)
    {
        graph.Add(new List<int>{ u, v, w });
    }
 
    // The main function to construct MST using Kruskal's
    // algorithm
    public void boruvkaMST()
    {
        List<int> parent = new List<int>();
 
        // An array to store index of the cheapest edge of
        // subset. It store [u,v,w] for each component
        List<int> rank = new List<int>();
        List<List<int> > cheapest = new List<List<int> >();
 
        // Initially there are V different trees.
        // Finally there will be one tree that will be MST
        int numTrees = V;
        int MSTweight = 0;
 
        // Create V subsets with single elements
        for (int node = 0; node < V; node++) {
            parent.Add(node);
            rank.Add(0);
            cheapest.Add(new List<int>{ -1, -1, -1 });
        }
 
        // Keep combining components (or sets) until all
        // components are not combined into single MST
        while (numTrees > 1) {
 
            // Traverse through all edges and update
            // cheapest of every component
            for (int i = 0; i < graph.Count; i++) {
 
                // Find components (or sets) of two corners
                // of current edge
                int u = graph[i][0], v = graph[i][1],
                    w = graph[i][2];
                int set1 = Find(parent, u),
                    set2 = Find(parent, v);
 
                // If two corners of current edge belong to
                // same set, ignore current edge. Else check
                // if current edge is closer to previous
                // cheapest edges of set1 and set2
                if (set1 != set2) {
                    if (cheapest[set1][2] == -1
                        || cheapest[set1][2] > w) {
                        cheapest[set1]
                            = new List<int>{ u, v, w };
                    }
                    if (cheapest[set2][2] == -1
                        || cheapest[set2][2] > w) {
                        cheapest[set2]
                            = new List<int>{ u, v, w };
                    }
                }
            }
 
            // Consider the above picked cheapest edges and
            // add them to MST
            for (int node = 0; node < V; node++) {
 
                // Check if cheapest for current set exists
                if (cheapest[node][2] != -1) {
                    int u = cheapest[node][0],
                        v = cheapest[node][1],
                        w = cheapest[node][2];
                    int set1 = Find(parent, u),
                        set2 = Find(parent, v);
                    if (set1 != set2) {
                        MSTweight += w;
                        UnionSet(parent, rank, set1, set2);
                        Console.WriteLine(
                            "Edge {0}-{1} with weight {2} included in MST",
                            u, v, w);
                        numTrees--;
                    }
                }
            }
            for (int node = 0; node < V; node++) {
                // reset cheapest array
                cheapest[node][2] = -1;
            }
        }
        Console.WriteLine("Weight of MST is {0}",
                          MSTweight);
    }
}
 
public class GFG {
    static void Main(string[] args)
    {
        Graph g = new Graph(4);
        g.AddEdge(0, 1, 10);
        g.AddEdge(0, 2, 6);
        g.AddEdge(0, 3, 5);
        g.AddEdge(1, 3, 15);
        g.AddEdge(2, 3, 4);
        g.boruvkaMST();
    }
}
// This code is contributed by prasad264

Javascript

// Boruvka's algorithm to find Minimum Spanning
// Tree of a given connected, undirected and weighted graph
 
// Class to represent a graph
class Graph {
    constructor(vertices) {
        this.V = vertices; // No. of vertices
        this.graph = []; // default dictionary to store graph
    }
 
    // function to add an edge to graph
    addEdge(u, v, w) {
        this.graph.push([u, v, w]);
    }
 
    // A utility function to find set of an element i
    // (uses path compression technique)
    find(parent, i) {
        if (parent[i] === i) {
            return i;
        }
        return this.find(parent, parent[i]);
    }
     
    // A function that does union of two sets of x and y
    // (uses union by rank)
    union(parent, rank, x, y) {
        const xroot = this.find(parent, x);
        const yroot = this.find(parent, y);
 
        // Attach smaller rank tree under root of high rank tree
        // (Union by Rank)
        if (rank[xroot] < rank[yroot]) {
            parent[xroot] = yroot;
        } else if (rank[xroot] > rank[yroot]) {
            parent[yroot] = xroot;
        }
         
        // If ranks are same, then make one as root and increment
        // its rank by one
        else {
            parent[yroot] = xroot;
            rank[xroot] += 1;
        }
    }
 
    //The main function to construct MST using Kruskal's algorithm
    boruvkaMST() {
        const parent = [];
         
        // An array to store index of the cheapest edge of
        // subset. It store [u,v,w] for each component
        const rank = [];
        const cheapest = [];
 
        // Initially there are V different trees.
        // Finally there will be one tree that will be MST
        let numTrees = this.V;
        let MSTweight = 0;
 
        // Create V subsets with single elements
        for (let node = 0; node < this.V; node++) {
            parent.push(node);
            rank.push(0);
            cheapest[node] = -1;
        }
 
        // Keep combining components (or sets) until all
        // components are not combined into single MST
        while (numTrees > 1) {
         
            // Traverse through all edges and update
            // cheapest of every component
            for (let i = 0; i < this.graph.length; i++) {
             
                // Find components (or sets) of two corners
                // of current edge
                const [u, v, w] = this.graph[i];
                const set1 = this.find(parent, u);
                const set2 = this.find(parent, v);
 
                // If two corners of current edge belong to
                // same set, ignore current edge. Else check if 
                // current edge is closer to previous
                // cheapest edges of set1 and set2
                if (set1 !== set2) {
                    if (cheapest[set1] === -1 || cheapest[set1][2] > w) {
                        cheapest[set1] = [u, v, w];
                    }
 
                    if (cheapest[set2] === -1 || cheapest[set2][2] > w) {
                        cheapest[set2] = [u, v, w];
                    }
                }
            }
 
            // Consider the above picked cheapest edges and add them
            // to MST
            for (let node = 0; node < this.V; node++) {
             
                // Check if cheapest for current set exists
                if (cheapest[node] !== -1) {
                    const [u, v, w] = cheapest[node];
                    const set1 = this.find(parent, u);
                    const set2 = this.find(parent, v);
 
                    if (set1 !== set2) {
                        MSTweight += w;
                        this.union(parent, rank, set1, set2);
                        console.log(`Edge ${u}-${v} with weight ${w} included in MST`);
                        numTrees--;
                    }
                }
            }
 
            for (let node = 0; node < this.V; node++) {
             
                // reset cheapest array
                cheapest[node] = -1;
            }
        }
 
        console.log(`Weight of MST is ${MSTweight}`);
    }
}
 
let g = new Graph(4);
g.addEdge(0, 1, 10);
g.addEdge(0, 2, 6);
g.addEdge(0, 3, 5);
g.addEdge(1, 3, 15);
g.addEdge(2, 3, 4);
 
g.boruvkaMST();
 
// This code is contributed by prajwal kandekar

Output

Edge 0-3 with weight 5 included in MST
Edge 0-1 with weight 10 included in MST
Edge 2-3 with weight 4 included in MST
Weight of MST is 19

Interesting Facts about Boruvka’s algorithm:

Time Complexity of Boruvka’s algorithm is O(E log V) which is the same as Kruskal’s and Prim’s algorithms.
Boruvka’s algorithm is used as a step in a faster randomized algorithm that works in linear time O(E).
Boruvka’s algorithm is the oldest minimum spanning tree algorithm that was discovered by Boruvka in 1926, long before computers even existed. The algorithm was published as a method of constructing an efficient electricity network.

Space complexity: The space complexity of Boruvka’s algorithm is O(V).

Exercise:
The above code assumes that the input graph is connected and it fails if a disconnected graph is given. Extend the above algorithm so that it works for a disconnected graph also and produces a forest.

Suggest improvement

Prim’s Algorithm for Minimum Spanning Tree (MST)

Dial's Algorithm (Optimized Dijkstra for small range weights)

Share your thoughts in the comments

Standard Greedy algorithms

Greedy algorithm on Array

Greedy algorithm on Operating System

Greedy algorithm on Graph

Approximate Greedy algorithm for NP complete problems

Greedy algorithm for special cases of DP

Some easy problems on Greedy algorithm

Some medium level problems on Greedy algorithm

Some hard problems on Greedy algorithm