How to initialise own values instead of using Tsp.create_random_instance

Question

In Qiskit's tutorial 'Max-Cut and Traveling Salesman Problem', below code is used to create the nodes:

tsp = Tsp.create_random_instance(n, seed=123)

What information is being generated by this function and how to replace this function if I have my own set of data?

Answer 1

The method Tsp.create_random_instance() creates a random instance of the TSP problem. Internally, it uses NetworkX's function random_geometric_graph().

To have your own graph, you can rely on NetworkX library as well to create it:

import networkx as nx

graph = nx.Graph()
graph.add_edges_from([[0, 2], [0, 3], [1, 2], [1, 3], [1, 4], [2, 4], [3, 4]])
nx.draw(graph, with_labels = True)

Now, pass the graph to Tsp constructor

tsp_instance = Tsp(graph)

Another method to create your own problem instance is by creating a TSPLIB file then use Tsp.parse_tsplib_format() method.