Paths formed by a noisy ant pheromone trail model

Abstract

The ant pheromone trail model is based on real ants depositing pheromones on their paths to indirectly communicate and create trails. These trails represent the shortest path from the ants' origin to their target. We apply this model to search for paths in a 2D lattice space and uniform lattice network with varying levels of noise. Without noise, a path is surely formed, while small noise leads to formation of self loops. Adding directionality with an additional pheromone helps path formation at low noise levels. We measured the paths formed and found them to be the shortest (normalized path length = 1.039). Paths formed with low noise persist longer than those with higher noise. In the lattice network, paths converge at a value greater than 1, whose convergence rate and value are influenced by the introduced noise. Ants may visit farther nodes with high pheromone intensity, preventing the formation of strictly shortest paths. Adjusting visibility and pheromone intensity between nodes is crucial for finding the shortest path in a network configuration.