ant colony optimization

Nov 8, 2011 at 3:05pm
Implementation of minimum spanning tree using ant colony optimization
in c or c++??????????????????????
Also tell me something ant colony optimization????
Nov 8, 2011 at 3:53pm
For something like this, probably C++.

As for ant colony optimization algorithms, I think the Wikipedia page describes it quite well, surprisingly.
http://en.wikipedia.org/wiki/Ant_colony_optimization_algorithms#Overview

-Albatross
Nov 8, 2011 at 4:17pm
You're going to have a very difficult time implementing an AOC without proper understanding of classes.
Nov 8, 2011 at 4:50pm
Oh, this isn't by any means a trivial problem. You'll need a fair understanding of classes, spanning trees and by extension graphs, the algorithm driving your "ants", and how to find a minimum spanning tree using them.

[Small hint: Put some restrictions on your ants.]

Good luck.

-Antbatross
Last edited on Nov 8, 2011 at 4:55pm
Nov 8, 2011 at 5:11pm
Wasn't there a P algorithm for minimum spanning trees though? Don't see much point to AOCing it...
Nov 8, 2011 at 5:54pm
Me neither, especially considering that the set of ant colony algorithms doesn't strike me as meant for purpose of finding minimum spanning trees. Shortest paths and cycles maybe. However, if it's an assignment and it specifically calls for an ant colony algorithm, then who are we to argue?

-Albatross
Last edited on Nov 8, 2011 at 5:54pm
Nov 9, 2011 at 1:34pm
I have to prepare a report on ant colony optimization in 2 or 3 days.please tell me more about ant colony optimization in a broader way except wikipedia????
Nov 9, 2011 at 3:52pm
Alright, alright...


An ACO algorithm is a probabilistic convergent algorithm in which you have a graph G, an arbitrary number of iterators "ants".

The edges of G have both weights and "pheromone" levels. The vertices can be normal, origin, or goal nodes.

As the "ants" travel the graph more or less randomly, they do not increase the "pheromone" levels. If they find a goal node, they will attempt to get back to their origin nodes, increasing the "pheromone" levels of the edges they take.

These "pheromones" bias the other "ants" to prefer that route over another of a lower "pheromone" level. If these "pheromones" are not reinforced by other "ants" taking the same route, they will decrease over time.

Eventually, the "ants" will mostly be traveling the shortest path from an origin node to a goal node.



Does this help? :)

-Albatross
Last edited on Nov 9, 2011 at 3:53pm
Nov 10, 2011 at 12:12pm
thanks Albatross............
Topic archived. No new replies allowed.