Open In App

Find the probability of a state at a given time in a Markov chain | Set 1

Improve
Improve
Improve
Like Article
Like
Save Article
Save
Share
Report issue
Report

Given a Markov chain G, we have the find the probability of reaching the state F at time t = T if we start from state S at time t = 0.

A Markov chain is a random process consisting of various states and the probabilities of moving from one state to another. We can represent it using a directed graph where the nodes represent the states and the edges represent the probability of going from one node to another. It takes unit time to move from one node to another. The sum of the associated probabilities of the outgoing edges is one for every node.

Consider the given Markov Chain( G ) as shown in below image:  

Examples

Input : S = 1, F = 2, T = 1
Output : 0.23
We start at state 1 at t = 0, 
so there is a probability of 0.23 
that we reach state 2 at t = 1.

Input : S = 4, F = 2, T = 100
Output : 0.284992

We can use dynamic programming and depth-first search (DFS) to solve this problem, by taking the state and the time as the two DP variables. We can easily observe that the probability of going from state A to state B at time t is equal to the product of the probability of being at A at time t – 1 and the probability associated with the edge connecting A and B. Therefore the probability of being at B at time t is the sum of this quantity for all A adjacent to B. 

Below is the implementation of the above approach:  

C++




// C++ implementation of the above approach
#include <bits/stdc++.h>
using namespace std;
 
// Macro for vector of pair to store
// each node with edge
#define vp vector<pair<int, float> >
 
// Function to calculate the
// probability of reaching F
// at time T after starting
// from S
float findProbability(vector<vp>& G, int N,
                      int F, int S, int T)
{
    // Declaring the DP table
    vector<vector<float> > P(N + 1, vector<float>(T + 1, 0));
 
    // Probability of being at S
    // at t = 0 is 1.0
    P[S][0] = 1.0;
 
    // Filling the DP table
    // in bottom-up manner
    for (int i = 1; i <= T; ++i)
        for (int j = 1; j <= N; ++j)
            for (auto k : G[j])
                P[j][i] += k.second * P[k.first][i - 1];
 
    return P[F][T];
}
 
// Driver code
int main()
{
    // Adjacency list
    vector<vp> G(7);
 
    // Building the graph
    // The edges have been stored in the row
    // corresponding to their end-point
    G[1] = vp({ { 2, 0.09 } });
    G[2] = vp({ { 1, 0.23 }, { 6, 0.62 } });
    G[3] = vp({ { 2, 0.06 } });
    G[4] = vp({ { 1, 0.77 }, { 3, 0.63 } });
    G[5] = vp({ { 4, 0.65 }, { 6, 0.38 } });
    G[6] = vp({ { 2, 0.85 }, { 3, 0.37 }, { 4, 0.35 }, { 5, 1.0 } });
 
    // N is the number of states
    int N = 6;
 
    int S = 4, F = 2, T = 100;
 
    cout << "The probability of reaching " << F
         << " at time " << T << " \nafter starting from "
         << S << " is " << findProbability(G, N, F, S, T);
 
    return 0;
}


Java




// Java implementation of the above approach
import java.util.*;
 
class GFG
{
    static class pair
    {
        int first;
        double second;
 
        public pair(int first, double second)
        {
            this.first = first;
            this.second = second;
        }
    }
 
    // Function to calculate the
    // probability of reaching F
    // at time T after starting
    // from S
    static float findProbability(Vector<pair>[] G,
                        int N, int F, int S, int T)
    {
        // Declaring the DP table
        float[][] P = new float[N + 1][T + 1];
         
        // Probability of being at S
        // at t = 0 is 1.0
        P[S][0] = (float) 1.0;
 
        // Filling the DP table
        // in bottom-up manner
        for (int i = 1; i <= T; ++i)
            for (int j = 1; j <= N; ++j)
                for (pair k : G[j])
                    P[j][i] += k.second * P[k.first][i - 1];
 
        return P[F][T];
    }
 
    // Driver code
    public static void main(String[] args)
    {
        // Adjacency list
        Vector<pair>[] G = new Vector[7];
        for (int i = 0; i < 7; i++)
        {
            G[i] = new Vector<pair>();
        }
         
        // Building the graph
        // The edges have been stored in the row
        // corresponding to their end-point
        G[1].add(new pair(2, 0.09));
        G[2].add(new pair(1, 0.23));
        G[2].add(new pair(6, 0.62));
        G[3].add(new pair(2, 0.06));
        G[4].add(new pair(1, 0.77));
        G[4].add(new pair(3, 0.63));
        G[5].add(new pair(4, 0.65));
        G[5].add(new pair(6, 0.38));
        G[6].add(new pair(2, 0.85));
        G[6].add(new pair(3, 0.37));
        G[6].add(new pair(4, 0.35));
        G[6].add(new pair(5, 1.0));
 
        // N is the number of states
        int N = 6;
 
        int S = 4, F = 2, T = 100;
 
        System.out.print("The probability of reaching " + F +
                " at time " + T + " \nafter starting from " +
                S + " is "
                + findProbability(G, N, F, S, T));
    }
}
 
// This code is contributed by Rajput-Ji


Python3




# Python3 implementation of the above approach
  
# Macro for vector of pair to store
# each node with edge
# define vp vector<pair<int, float> >
  
# Function to calculate the
# probability of reaching F
# at time T after starting
# from S
def findProbability(G, N, F, S, T):
 
    # Declaring the DP table
    P = [[0 for i in range(T + 1)]
            for j in range(N + 1)]
  
    # Probability of being at S
    # at t = 0 is 1.0
    P[S][0] = 1.0;
  
    # Filling the DP table
    # in bottom-up manner
    for i in range(1, T + 1):
        for j in range(1, N + 1):
            for k in G[j]:
                P[j][i] += k[1] * P[k[0]][i - 1];
  
    return P[F][T]
 
# Driver code
if __name__=='__main__':
 
    # Adjacency list
    G = [0 for i in range(7)]
  
    # Building the graph
    # The edges have been stored in the row
    # corresponding to their end-point
    G[1] = [ [ 2, 0.09 ] ]
    G[2] = [ [ 1, 0.23 ], [ 6, 0.62 ] ]
    G[3] = [ [ 2, 0.06 ] ]
    G[4] = [ [ 1, 0.77 ], [ 3, 0.63 ] ]
    G[5] = [ [ 4, 0.65 ], [ 6, 0.38 ] ]
    G[6] = [ [ 2, 0.85 ], [ 3, 0.37 ],
             [ 4, 0.35 ], [ 5, 1.0 ] ]
  
    # N is the number of states
    N = 6
  
    S = 4
    F = 2
    T = 100
     
    print("The probability of reaching {} at "
          "time {}\nafter starting from {} is {}".format(
          F, T, S, findProbability(G, N, F, S, T)))
  
# This code is contributed by rutvik_56


C#




// C# implementation of the above approach
using System;
using System.Collections.Generic;
 
class GFG
{
    class pair
    {
        public int first;
        public double second;
 
        public pair(int first, double second)
        {
            this.first = first;
            this.second = second;
        }
    }
 
    // Function to calculate the
    // probability of reaching F
    // at time T after starting
    // from S
    static float findProbability(List<pair>[] G,
                        int N, int F, int S, int T)
    {
        // Declaring the DP table
        float[,] P = new float[N + 1, T + 1];
         
        // Probability of being at S
        // at t = 0 is 1.0
        P[S, 0] = (float) 1.0;
 
        // Filling the DP table
        // in bottom-up manner
        for (int i = 1; i <= T; ++i)
            for (int j = 1; j <= N; ++j)
                foreach (pair k in G[j])
                    P[j, i] += (float)k.second *
                                P[k.first, i - 1];
 
        return P[F, T];
    }
 
    // Driver code
    public static void Main(String[] args)
    {
        // Adjacency list
        List<pair>[] G = new List<pair>[7];
        for (int i = 0; i < 7; i++)
        {
            G[i] = new List<pair>();
        }
         
        // Building the graph
        // The edges have been stored in the row
        // corresponding to their end-point
        G[1].Add(new pair(2, 0.09));
        G[2].Add(new pair(1, 0.23));
        G[2].Add(new pair(6, 0.62));
        G[3].Add(new pair(2, 0.06));
        G[4].Add(new pair(1, 0.77));
        G[4].Add(new pair(3, 0.63));
        G[5].Add(new pair(4, 0.65));
        G[5].Add(new pair(6, 0.38));
        G[6].Add(new pair(2, 0.85));
        G[6].Add(new pair(3, 0.37));
        G[6].Add(new pair(4, 0.35));
        G[6].Add(new pair(5, 1.0));
 
        // N is the number of states
        int N = 6;
 
        int S = 4, F = 2, T = 100;
 
        Console.Write("The probability of reaching " + F +
                " at time " + T + " \nafter starting from " +
                S + " is "
                + findProbability(G, N, F, S, T));
    }
}
 
// This code is contributed by 29AjayKumar


Javascript




<script>
 
// Javascript implementation of the above approach
 
// Function to calculate the
// probability of reaching F
// at time T after starting
// from S
function findProbability(G, N, F, S, T)
{
    // Declaring the DP table
    var P = Array.from(Array(N+1), ()=> Array(T+1).fill(0))
 
    // Probability of being at S
    // at t = 0 is 1.0
    P[S][0] = 1.0;
 
    // Filling the DP table
    // in bottom-up manner
    for (var i = 1; i <= T; ++i)
        for (var j = 1; j <= N; ++j)
            G[j].forEach(k => {
                 
                P[j][i] += k[1] * P[k[0]][i - 1];
            });
 
    return P[F][T];
}
 
// Driver code
// Adjacency list
var G = Array(7);
// Building the graph
// The edges have been stored in the row
// corresponding to their end-point
G[1] = [ [ 2, 0.09 ] ];
G[2] = [ [ 1, 0.23 ], [ 6, 0.62 ] ];
G[3] = [ [ 2, 0.06 ] ];
G[4] = [ [ 1, 0.77 ], [ 3, 0.63 ] ];
G[5] = [ [ 4, 0.65 ], [ 6, 0.38 ] ];
G[6] = [ [ 2, 0.85 ], [ 3, 0.37 ], [ 4, 0.35 ], [ 5, 1.0 ] ];
// N is the number of states
var N = 6;
var S = 4, F = 2, T = 100;
document.write( "The probability of reaching " + F
     + " at time " + T + " <br>after starting from "
     + S + " is " + findProbability(G, N, F, S, T).toFixed(8));
 
 
</script>


Output: 

The probability of reaching 2 at time 100 
after starting from 4 is 0.284992

 

Complexity Analysis:

  • Time complexity: O(N2 * T) 
  • Space complexity: O(N * T)


Last Updated : 06 Sep, 2022
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads