A practical application – solving the TSP

Let’s first look at the problem statement for the TSP, which is a well-known problem that was coined as a challenge in the 1930s. The TSP is an NP-hard problem. To start with, we can randomly generate a tour that meets the condition of visiting all of the cities without caring about the optimal solution. Then, we can work to improve the solution with each iteration. Each tour generated in an iteration is called a candidate solution (also called a certificate). Proving that a certificate is optimal requires an exponentially increasing amount of time. Instead, different heuristics-based solutions are used that generate tours that are near to optimal but are not optimal.

A traveling salesperson needs to visit a given list of cities to get their job done:

INPUT

A list of n cities (denoted as V) and the distances between each pair of cities, d ij (1...