hungarian-on3

0.3.1 • Public • Published

hungarian-on3

The hungarian (Kuhn-Munkres) algorithm solved in O(n^3) time Algorithm based on: https://github.com/KevinStern/software-and-algorithms/blob/master/src/main/java/blogspot/software_and_algorithms/stern_library/optimization/HungarianAlgorithm.java

Solves an assignment problem really fast. Benchmarked against other JS solutions, a 1000x1000 matrix took about 26 seconds. With this, it takes 2 seconds.

  • The primary speed up comes from using the O(n^3) algorithm instead of the O(n^4) algorithm that you probably use when you solve by hand.
  • Additionally, it reduces the search space by removing rows that have no feasible column match.
  • For max speed, make sure your BIG_M is <= 2^30 (otherwise the optimizing compiler bails out because it has to rewrap the function in a double)

Installation

npm install hungarian-on3

How to use

var hungarian = require('./hungarian-on3');
var data = [[400, 150, 400],[400, 450, 600],[300, 225, 300]];
var results = hungarian(data);
//results: [[0, 1], [1, 0], [2, 2]]

License

MIT

Package Sidebar

Install

npm i hungarian-on3

Weekly Downloads

6

Version

0.3.1

License

MIT

Last publish

Collaborators

  • mattkrick