PyPOPMusic

What is it?

PyPOPMusic is a simple Python wrapper around the POPMusic TSP solver implemented in LKH 2.0.9.

PyPOPMusic allows you to compute candidate lists for the Traveling Salesman Problem. It uses the fast LKH solver implemented in C and provides a convenient Python layer around it.

How do I install it?

PyPOPMusic is tested under Python 3.8. It contains a modified version of the C effective implementation of the Lin-Kernighan heuristic for solving the traveling salesman problem.

To build PyPOPMusic, clone the repository:

    git clone https://github.com/tommivitali/pypopmusic
    cd pypopmusic

Then run

    pip install -e .

to install all the dependencies required to use PyPOPMusic, and to build the modified version of the LKH C implementation.

What can I do with it?

The main entry point of the PyPOPMusic library is the PyCandidatePOP class. Here we create a random set of points, and use it to create all the candidate lists.

    >>> import numpy as np
    >>> from pypopmusic.PyCandidatePOP import PyCandidatePOP
    >>> n_cities = np.random.randint(100, 300)
    >>> pos = np.random.uniform(-0.5, 0.5, size=n_cities * 2).reshape((n_cities, 2))
    >>> instance_popm = PyCandidatePOP()
    >>> cand_lists, time_spent = instance_popm(pos)
    Lower bound = 1256566.8, Ascent time = 0.23 sec.
    Cand.min = 2, Cand.avg = 5.3, Cand.max = 9
    Preprocessing time = 0.23 sec.
    Total execution time: 0.2389 seconds

As you can see above, PyPOPMusic is somewhat chatty and will print various message to the standard output. You can avoid these messages to be printed specifying the verbose=False parameter in the constructor.