deanone/dijkstra

Native Python implementation of Dijkstra's algorithm.

Native Python implementation of Dijkstra's algorithm.

https://en.wikipedia.org/wiki/Dijkstra%27s_algorithm

This implementation (at its current phase) does not include a priority queue.

Usage:

1. Directly from terminal: python dijkstra.py A B C

2. From ipython: %run dijkstra.py A B C

Arguments:

1. A: This is the execution mode. It can take 2 values, namely 0 which means that the script will use as graph the one stored in graph.csv file, and 1 which means that a random graph (in the form of a random symmetric numpy array of integers) will be used.

2. B: the source node (as integer)

3. C: the target node (as integer)

Example: for a graph with N = 10 nodes stored in graph.csv file, source = 0 and target = 5 the execution command will be (from within ipython):

%run dijkstra.py 0 0 5

Performance:

So far the algorithm was tested on graphs with different sizes, with the largest one having 10000 nodes and the corresponding adjacency matrix having 10^8 values. For this graph, the execution time was approximately 47 seconds. It's pretty slow right now but it will be faster :-)