Skip to content
Related Articles

Related Articles

Maximize shortest path between given vertices by adding a single edge
  • Last Updated : 13 Jan, 2021
GeeksforGeeks - Summer Carnival Banner

Given an undirected graph of N nodes and M vertices. You are also given a K edges as selected[]. The task to maximize the shortest path length between node 1 to node N by adding single edges between any two vertices from the given selected edges. 
Note: You may add an edge between any two selected vertices who have already an edge between them.

Input: N = 5, M = 4, K = 2, selected[] = {2, 4} 
Below is the given graph: 
 

Output:
Explanation: 
Before adding an edge between 2 and 4, the Shortest Path becomes: 1–>2–>3–>4–>5. 
After adding an edge between 2 and 4, the Shortest Path becomes 1–>2–>4–>5. Below is the graph after adding edges. denoted by the dashed line. 
 



Input: N = 5 M = 5 K = 3 selected[] = {1, 3, 5} 
Below is the given graph: 
 

Output:
Explanation: 
We can add an edge between 3 and 5 as they have already an edge between them. so, the shortest path becomes 1–>2–>3–>5. Below is the graph after adding edges. denoted by the dashed line. 
 

 

 

Approach: The idea is to use Breadth-First Search to find the distance from vertices 1 and N to each selected vertex. For selected vertex i, let xi denote the distance to node 1, and yi denote the distance to node N. Below are the steps:

  1. Maintain a 2D matrix(say dist[2][]) having 2 rows and N columns.
  2. In the first row, Maintain the shortest distance between node 1 and other vertices in the graph using BFS Traversal.
  3. In the second row, Maintain the shortest distance between node N and the other vertices of the graph using BFS Traversal.
  4. Now, choose two selected vertices a and b from selected[] to minimize the value of min(xa + yb, ya + xb). For this do the following: 
    • Create a vector of pairs and store the value of (xi – yi) with their respective selected node.
    • Sort the above vector of pairs.
    • Initialize best to 0 and max to -INF.
    • Now traverse the above vector of pairs and for each selected node(say a) update the value of best to maximum of (best, max + dist[1][a]) and update max to maximum of (max, dist[0][a]).
  5. After the above operations the maximum of (dist[0][N-1] and best + 1) given the minimum shortest path.

Below is the implementation of the above approach:

C++




// C++ program for the above approach
#include <bits/stdc++.h>
using namespace std;
const int INF = 1e9 + 7;
int N, M;
 
// To store graph as adjacency list
vector<int> edges[200005];
 
// To store the shortest path
int dist[2][200000];
 
// Function that performs BFS Traversal
void bfs(int* dist, int s)
{
    int q[200000];
 
    // Fill initially each distance as INF
    fill(dist, dist + N, INF);
    int qh = 0, qt = 0;
    q[qh++] = s;
    dist[s] = 0;
 
    // Perform BFS
    while (qt < qh) {
 
        int x = q[qt++];
 
        // Traverse the current edges
        for (int y : edges[x]) {
            if (dist[y] == INF) {
 
                // Update the distance
                dist[y] = dist[x] + 1;
 
                // Insert in queue
                q[qh++] = y;
            }
        }
    }
}
 
// Function that maximizes the shortest
// path between source and destination
// vertex by adding a single edge between
// given selected nodes
void shortestPathCost(int selected[], int K)
{
    vector<pair<int, int> > data;
 
    // To update the shortest distance
    // between node 1 to other vertices
    bfs(dist[0], 0);
 
    // To update the shortest distance
    // between node N to other vertices
    bfs(dist[1], N - 1);
 
    for (int i = 0; i < K; i++) {
 
        // Store the values x[i] - y[i]
        data.emplace_back(dist[0][selected[i]]
                              - dist[1][selected[i]],
                          selected[i]);
    }
 
    // Sort all the vectors of pairs
    sort(data.begin(), data.end());
    int best = 0;
    int MAX = -INF;
 
    // Traverse data[]
    for (auto it : data) {
        int a = it.second;
        best = max(best,
                   MAX + dist[1][a]);
 
        // Maximize x[a] - y[b]
        MAX= max(MAX, dist[0][a]);
    }
 
    // Print minimum cost
    printf("%d\n", min(dist[0][N - 1], best + 1));
}
 
// Driver Code
int main()
{
    // Given nodes and edges
    N = 5, M = 4;
    int K = 2;
    int selected[] = { 1, 3 };
 
    // Sort the selected nodes
    sort(selected, selected + K);
 
    // Given edges
    edges[0].push_back(1);
    edges[1].push_back(0);
    edges[1].push_back(2);
    edges[2].push_back(1);
    edges[2].push_back(3);
    edges[3].push_back(2);
    edges[3].push_back(4);
    edges[4].push_back(3);
 
    // Function Call
    shortestPathCost(selected, K);
    return 0;
}

Java




// Java program for the above approach
import java.util.*;
import java.lang.*;
 
class GFG{
     
static int INF = (int)1e9 + 7;
static int N, M;
   
// To store graph as adjacency list
static ArrayList<ArrayList<Integer>> edges;
   
// To store the shortest path
static int[][] dist = new int[2][200000];
   
// Function that performs BFS Traversal
static void bfs(int[] dist, int s)
{
    int[] q = new int[200000];
   
    // Fill initially each distance as INF
    Arrays.fill(dist, INF);
     
    int qh = 0, qt = 0;
    q[qh++] = s;
    dist[s] = 0;
   
    // Perform BFS
    while (qt < qh)
    {
        int x = q[qt++];
   
        // Traverse the current edges
        for(Integer y : edges.get(x))
        {
            if (dist[y] == INF)
            {
                 
                // Update the distance
                dist[y] = dist[x] + 1;
   
                // Insert in queue
                q[qh++] = y;
            }
        }
    }
}
   
// Function that maximizes the shortest
// path between source and destination
// vertex by adding a single edge between
// given selected nodes
static void shortestPathCost(int selected[], int K)
{
    ArrayList<int[]> data = new ArrayList<>();
   
    // To update the shortest distance
    // between node 1 to other vertices
    bfs(dist[0], 0);
   
    // To update the shortest distance
    // between node N to other vertices
    bfs(dist[1], N - 1);
   
    for(int i = 0; i < K; i++)
    {
         
        // Store the values x[i] - y[i]
        data.add(new int[]{dist[0][selected[i]] -
                           dist[1][selected[i]],
                                   selected[i]});
    }
   
    // Sort all the vectors of pairs
    Collections.sort(data, (a, b) -> a[0] - b[0]);
    int best = 0;
    int MAX = -INF;
   
    // Traverse data[]
    for(int[] it : data)
    {
        int a = it[1];
        best = Math.max(best,
                        MAX + dist[1][a]);
   
        // Maximize x[a] - y[b]
        MAX = Math.max(MAX, dist[0][a]);
    }
     
    // Print minimum cost
    System.out.println(Math.min(dist[0][N - 1],
                                     best + 1));
}
 
// Driver code
public static void main (String[] args)
{
     
    // Given nodes and edges
    N = 5; M = 4;
    int K = 2;
    int selected[] = { 1, 3 };
     
    // Sort the selected nodes
    Arrays.sort(selected);
     
    edges = new ArrayList<>();
     
    for(int i = 0; i < 200005; i++)
        edges.add(new ArrayList<Integer>());
     
    // Given edges
    edges.get(0).add(1);
    edges.get(1).add(0);
    edges.get(1).add(2);
    edges.get(2).add(1);
    edges.get(2).add(3);
    edges.get(3).add(2);
    edges.get(3).add(4);
    edges.get(4).add(3);
     
    // Function Call
    shortestPathCost(selected, K);
}
}
 
// This code is contributed by offbeat

Python3




# Python3 program for the above approach
 
# Function that performs BFS Traversal
def bfs(x, s):
    global edges, dist
    q = [0 for i in range(200000)]
 
    # Fill initially each distance as INF
    # fill(dist, dist + N, INF)
    qh, qt = 0, 0
    q[qh] = s
    qh += 1
    dist[x][s] = 0
 
    # Perform BFS
    while (qt < qh):
        xx = q[qt]
        qt += 1
 
        # Traverse the current edges
        for y in edges[xx]:
            if (dist[x][y] == 10**18):
 
                # Update the distance
                dist[x][y] = dist[x][xx] + 1
 
                # Insert in queue
                q[qh] = y
                qh += 1
 
# Function that maximizes the shortest
# path between source and destination
# vertex by adding a single edge between
# given selected nodes
def shortestPathCost(selected, K):
    global dist, edges
    data = []
 
    # To update the shortest distance
    # between node 1 to other vertices
    bfs(0, 0)
 
    # To update the shortest distance
    # between node N to other vertices
    bfs(1, N - 1)
    for i in range(K):
 
        # Store the values x[i] - y[i]
        data.append([dist[0][selected[i]]- dist[1][selected[i]], selected[i]])
 
    # Sort all the vectors of pairs
    data = sorted(data)
    best = 0
    MAX = -10**18
 
    # Traverse data[]
    for it in data:
        a = it[1]
        best = max(best,MAX + dist[1][a])
 
        # Maximize x[a] - y[b]
        MAX= max(MAX, dist[0][a])
 
    # Prminimum cost
    print(min(dist[0][N - 1], best + 1))
 
# Driver Code
if __name__ == '__main__':
 
    # Given nodes and edges
    edges = [[] for i in range(5)]
    dist = [[10**18 for i in range(1000005)] for i in range(2)]
    N,M = 5, 4
    K = 2
    selected = [1, 3]
 
    # Sort the selected nodes
    selected = sorted(selected)
 
    # Given edges
    edges[0].append(1)
    edges[1].append(0)
    edges[1].append(2)
    edges[2].append(1)
    edges[2].append(3)
    edges[3].append(2)
    edges[3].append(4)
    edges[4].append(3)
 
    # Function Call
    shortestPathCost(selected, K)
 
    # This code is contributed by mohit kumar 29
Output: 
3

 

Time Complexity: O(N*log N + M) 
Auxiliary Space: O(N) 
 

 

Attention reader! Don’t stop learning now. Get hold of all the important DSA concepts with the DSA Self Paced Course at a student-friendly price and become industry ready.

My Personal Notes arrow_drop_up
Recommended Articles
Page :