Design and Analysis of Algorithms
Dynamic Programming
(Travelling Salesman Problem)
Dr. D. P. Acharjya
Professor, SCOPE
Office: SJT Annex 201E
Email: [email protected]
Dr. D. P. Acharjya 2/21/2024 1
Travelling Salesman Problem
The job of a travelling salesman is to visit all the cities
linked by roads in a particular territory.
He has to visit all the cities exactly once and to return
his home city in such a manner that, the total cost
(distance travelled, time, money) is minimum.
The above problem can be visualized as a graph theory
problem. Consider all the cities as vertices and the
roads connected to them as edges.
So, it is a weighted directed graph.
Our objective is to find a minimum cost Hamiltonian
tour for a graph.
Dr. D. P. Acharjya 2/21/2024 2
Complete Graph, Walk and Tour
A graph G(V, E); E (VV) is said to be a
complete graph, if for every pair of vertices, there
exists an edge.
A walk is a sequence of vertices and edges of a
graph G(V, E).
A tour is a walk in which the start and end vertex is
same.
A tour in which all the edges are travelled exactly
once is known as Eulerian tour.
A tour in which all the vertices are visited exactly
once is known as Hamiltonian tour.
Dr. D. P. Acharjya 2/21/2024 3
Mathematical Formulation
Let G(V,E) be a complete directed graph. Let the cost from
ith vertex to jth vertex be cij.
Let |V| = n such that n > 1.
Let us consider the tour starts from vertex 1 and ends at
vertex 1.
Let g(i, S) be the length of a shortest path starting at vertex
i going through all vertices in S and terminating at vertex i.
Dr. D. P. Acharjya 2/21/2024 4
Continued …
Above equation is executed for |S| = 1, 2, 3, … with i 1
and i S.
If |S| = (n - 1), then the optimal tour is defined as:
Dr. D. P. Acharjya 2/21/2024 5
Numerical Illustration
Consider the cities are connected as per the
following graph. Find an optimal tour to the
travelling salesman problem over the given graph.
Dr. D. P. Acharjya 2/21/2024 6
Solution
The adjacency matrix of the above graph with
respect to the vertices 1, 2, 3, and 4 is given as:
Let the source vertex be 1. Thus we get:
Dr. D. P. Acharjya 2/21/2024 7
Continued …
For |S| = 1, we have S = {3} or {4} for i = 2.
Dr. D. P. Acharjya 2/21/2024 8
Continued …
For |S| = 2, we have S = {3, 4} for i = 2.
Dr. D. P. Acharjya 2/21/2024 9
Continued …
For |S| = 2, we have S = {2, 3} for i = 4.
Dr. D. P. Acharjya 2/21/2024 10
Continued …
The optimal tour is given as |S| = (n-1) = 3
The Optimal Tour is given below with cost = 35
12431
Dr. D. P. Acharjya 2/21/2024 11
Computing Time
The total number of computations for g(i, S) for |S|
= k is given as:
The number of comparisons in optimal tour is (n -
2) as |S| = (n - 1)
Total time is given as:
Dr. D. P. Acharjya 2/21/2024 12
