# pyEntropy/elkai

Python 3 TSP solver based on LKH (cross platform)
Latest commit c59afe9 May 18, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
LKH-2.0.9 May 4, 2019
elkai May 14, 2019
.gitignore May 3, 2019
.travis.yml May 14, 2019
CMakeLists.txt May 3, 2019
appveyor.yml May 14, 2019
build_wheels.sh May 14, 2019
dev-requirements.txt May 3, 2019
setup.py May 14, 2019

# elkai - a Python 3 TSP solver

elkai is a Python 3 library for solving travelling salesman problems without external dependencies, based on LKH by Keld Helsgaun.

💾 To install it run `pip install elkai`

💻 Supported platforms: elkai is available on Windows, Linux, OS X for Python 3.5 and above as a binary wheel.

## Example usage

```import elkai
elkai.solve_int_matrix([
[1, 1],
[1, 1]
]) # Output: [0, 1]```

## Documentation

solve_int_matrix(matrix, runs=10) solves an instance of the asymmetric TSP problem with integer distances.

• `matrix`: an N*N matrix representing an integer distance matrix between cities.

An example of N=3 city arrangement:

```[                  # cities are zero indexed, d() is distance
[0, 4,  9],    # d(0, 0), d(0, 1), d(0, 2)
[4, 0, 10],    # d(1, 0), d(1, 1), ...
[2, 4,  0]     # ... and so on
]```
• `runs`: An integer representing how many iterations the solver should perform. By default, 10.

• Return value: The tour represented as a list of indices. The indices are zero-indexed and based on the distance matrix order.

solve_float_matrix(matrix, runs=10) has the same signature as the previous, but allows floating point distances. It may be inaccurate.

## FAQ

What's the difference between LKH and elkai?

elkai is a library that contains the LKH solver and has some wrapper code to expose it to Python. The advantage is that you don't have to compile LKH yourself or download its executables and then manually parse the output. Note that elkai and its author are not affiliated with the LKH project. Note: Helsgaun released the LKH project for non-commercial use only, so elkai must be used this way too.

How to manually build the library?

You need CMake, a C compiler and Python 3.5+. You need to install the dev dependencies first: `pip install scikit-build ninja`. To build the library, run `python setup.py build` and `python setup.py install` to install it. To make a binary wheel, run `python setup.py bdist_wheel`.

How accurately does it solve asymmetric TSP problems?

Instances with known solutions, which are up to N=315 cities, can be solved optimally.