In nature, ants forage for food using a phenomenon called stigmergy, whereby the ants indirectly communicate by placing pheromone in the environment for other ants to find. To test whether this strategy is optimal, a functional evolutionary algorithm system for improving the behavior of pheromone-dropping ”ant” agents in simulation is presented. Traditional RL methods are tested as alternatives, but no neural network agent is able to approach the performance of a simple pre-programmed agent, suggesting that the foraging problem is fundamentally difficult to train a neural network decision agent for.
elijah-rou/Ant-Neuroevolution
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|