Small Feedback Arc Set (sfas)

Efficient implementation of a greedy algorithm for computing small feedback arc sets in directed weighted multi-graphs. This implementation is an adaptation of the algorithm described in Section 2.3 of this article, with additional generalization to support weights and parallel edges.

Description

Given a weighted directed graph, computes a topological sorting (linear order of the nodes) that minimizes (greedily) the number of backward edges - feedback arc set. In particular, removing the set of edges going backward in the resulting order breaks all directed cycles in the graph.

Install

pip install sfas

Example usage

graph = [
    ['a', 'b', 1],
    ['a', 'c', 1],
    ['c', 'd', 2],
    ['d', 'a', 2],
]

from sfas import greedy
ouput = greedy.compute_order(graph)

output

['c', 'd', 'a', 'b']

Interface

Params:

1. connections : List[List[Any, Any, Int]] - list of edges, each represented as a 3-item list consisting of [<from node>, <to node>, <edge weight>]
2. verbosity : Int - prints progress and other stats for values > 0
3. random_seed : Int randomness is in picking the next "greedy" step among equally qualified ones

Result:

1. List with all nodes, ordered so that the total weight of edges going backwards (w.r.t. this order) is small

Questions / suggestions welcome

arie.matsliah@gmail.com