A simulation of ant formation and foraging using fuzzy logic and Reinforcement Learning

Pheromone trails laid by foraging ants serve as a positive feedback mechanism in the ant colonies to share information (in search of food sources). The simulation conducted here of this swarm intelligence can help to realize the process and implement it further for artificial swarms. With available instrumentation we may easily record the agentspsila position in each step. Pheromone trails are then generated and each agent learns to follow the trail. Fuzzy logic is used to approximate pheromone value at each point. Agents learn to behave like ants using a reinforcement learning (RL) process. The algorithm has appropriate parameters that may be set according to the search space dimension. Simulation results are in good agreement with the observations of antspsila behavior.