Package acogo provides a framework for playing with ant colony optimization.

acogo will run a set number of ants from start to home in the graph and then output a DOT language graph of the resulting graph pheromone levels.

godep go build

To write the output DOT graph to stdout, run

./acogo

To transform the output DOT graph to a png, run

./acogo | dot -Tpng -o antgraph.png

-antcount: The number of ants to create and run each iteration. Default 20.
-depositamt: The amount of pheromone each ant deposits on their path. Default 1.0.
-iterations: The number of times each ant runs from home to goal. Default 500.
-decay: The amount that pheromone on each edge decreases after each iteration. Default 0.3.
-dimension: The number of nodes on each side of the square graph. Default 6.
-start: The index of the node from which the ants start. Default 0.
-goal: The index of the node that ants are trying to reach. Default dimension * dimension.

When run, acogo will create a square graph of size dimension * dimension with each node having connections to adjacent nodes above, below, left, right, and on all four diagonals. For example, if dimension is 3, the graph look like:

0 1 2
3 4 5
6 7 8

antcount ants will be placed at the start node on the graph and travel in the graph until they reach the goal node. At each node, the ant will probabilistically chose which node to travel to next proportionate to the amount of pheromone on each outgoing node. Ants will not return to the node they just left unless that is the only option for exiting a particular node.

Once all ants reach the goal node in a given iteration, depositamt pheromone will be added to each edge that each ant traveled on. After reaching the goal, ants routes are unlooped, so an ant that traveled 1->4->5->2->4->8 would only lay down pheromone from 1->4->8. After all ant pheromone has been laid down, decay pheromone is subtracted from all edges.

When all ants have completed iterations iterations, a DOT language representation of the final graph is written to stdout. Edge colors reflect how much pheromone is on a given edge with darker edges representing more pheromone.