Question: Ant Colony Optimization (ACO) algorithms are inspired by the collective trail-laying, trail following foraging behavior of ants. Individual ants deposit pheromone as they move from a food source to their nest and foragers follow such pheromone trails. An ACO algorithm involves the following main steps:
• A probabilistic transition rule that defines the next step of an "artificial ant". This is a function of the pheromone trail.
• After completing a tour, an ant lays some pheromone trail on the tour it performed. This is inversely proportional to the length of the tour performed.
• The pheromone trail evaporates with time.
Consider the following three rules that could be implemented to create an ALife simulation of ant foraging:
• If an ant does not know where there is food, then forage.
• If an ant sees another ant, ask if it knows where there is food. If it does, go for it. Else, run away from that ant.
• If an ant finds food, it must take it back to the nest, and return to the place where it found food. If an ant does not find any food where it was supposed to be, then keep foraging
Explain the behaviors that would emerge from this rule set.
Discuss what would probably happen if there is a single food source on the grass, and if there are more than one food source.
Do you think the ants would privilege food sources that are nearer to the nest? If yes, why? If not, should they?