This python library gives a solver to efficiently find good sub-optimal solutions to the TSP. It uses the 2-opt heuristic and the search part is implemented in rust.

To install the library, run:

pip install tsp_2opt

For now, the library is working only on MacOS X platforms.

The current version relies on the availability of a distance matrix, i.e. matrix giving pair-wise distance between cities in the TSP instance.

Here is a small example :

from tsp_2opt.solver import tsp_solver
distances = [[0, 2, 3], [2, 0, 4], [3, 4, 0]]
route, length = tsp_solver(distances)

The solution obtained to a TSP instance after performing the 2-opt heuristic depends on the initial solution (starting point of the search), therefore this library runs the heuristic n times with n different starting point in parralel (true parallelism is possible thanks to rust) and the best obtained is returned. You can pass n as a parameter to tsp_solver() and control the maximum number of workers :

tsp_solver(distances, workers, n)