Overview of SWARM INTELLIGENCE and ANT COLONY OPTIMIZATION

Overview of SWARM INTELLIGENCE and ANT COLONY OPTIMIZATION

SWARM INTELLIGENCE � Based on social interactions (locally shared knowledge) that provides the basis for unguided problem solving. � Efficiency is related to the degree of connectedness of the network and the number of interacting agents.

CHARACTERISTICS OF SWARM � Distributed, no central control � Limited communication � No explicit model of environment � Perception of the environment � Composed of many, alike individual agents.

Examples � Ant colony optimization � River formation dynamics � Particle swarm optimization � Gravitational search algorithm � Intelligent water drops

ANT COLONY OPTIMIZATION � Developed by M. Dorgio in 1992 � Heuristic optimization method inspired by the observation of real ant colonies. � Based on how ants find the shortest path to food source. � The behavior of ants is a kind of stochastic distributed optimization behavior.

BEHAVIOR OF REAL ANTS � Ants are blind, deaf and dumb. � So how do they find the shortest path to food sources? ◦ Based on PHEROMONES. ◦ They follow the deposits of pheromones and form a trail. ◦ Other ants get attracted to this trail. � Pheromones are volatile in nature.

CONTD… � Each ant choose an action based on ◦ Random choice ◦ Pheromone mediated � They move by sensing previous ant not by sensing the environment. � Each ant collects info about local environment and act concurrently and independently. � Stigmergy governs info exchange.

APPLICATIONS � Network routing � Travelling sales man problem � Vehicle routing � Assignment problems � Set problems