2OutlineOutline

Swarm Intelligence

Introduction to Ant Colony Optimization (ACO)

Ant Behaviour

Stigmergy

Pheromones

Basic Algorithm

Example

Advantages and Disadvantages

References

3Swarm IntelligenceSwarm Intelligence

Artificial intelligence technique based on the study of collective behavior in decentralized, self-organized systems

Introduced by Beni & Wang in 1989

Collective system capable of performing complex tasks in a dynamic environment

Model suited to distributed problem solving

Works without:External guidance Central coordination

Typically made up of a population of simple agents

5Ant Colony OptimizationAnt Colony Optimization

Proposed by Marco Dorigo in 1991

Inspired in the behavior of real ants

Multi-agent approach for solving complex combinatorial optimization problems

Applications:

Traveling Salesman ProblemSchedulingNetwork Model ProblemVehicle routing

9Ant BehaviorAnt Behavior

Walk randomly

Lay pheromone Trail

Follow Trail

Found food

Start

Continue

Quest for food

10

Ant BehaviorAnt Behavior

Ant behavior is stochastic

The behavior is induced by indirect communication (pheromone paths) - Stigmergy

Ants explore the search space

Limited ability to sense local environment

Act concurrently and independently

High quality solutions emerge via global cooperation

11

Stigmergy

Term coined by French biologist Pierre-Paul Grasse, means interaction through the environment

Indirect communication via interaction with environment

Agents respond to changes in the environment

Allows simpler agents

Decreases direct communication

12

Pheromones

Ants lay pheromone trails while traveling

Pheromones accumulate with multiple ants using a path

This behavior leads to the appearance of shortest paths

Pheromones evaporateAvoids being trapped in local optima

small low evaporation slow adaptation

large high evaporation fast adaptation

Pheromones = long-term memory of an ant colony

13

ACO Algorithm

Construct solutions Explore the search spaceChoose next step probabilistically according to the

pheromone modelApply local search to constructed solutions (Optional) Update pheromones (add new + evaporate)

14

Example: TSP

15

Example: TSP

16

Advantages and Disadvantages

AdvantagesCan be used in dynamic applicationsPositive Feedback leads to rapid discovery of good

solutionsDistributed computation avoids

premature convergence

DisadvantagesConvergence is guaranteed, but time to convergence

uncertainCoding is not straightforward

17

References

Dorigo M., Blum C., Ant colony optimization theory: A survey, Theoretical Computer Science, Volume 344, Issues 2-3, November 2005Blum C., Ant colony optimization: Introduction and recent trends, Physics of Life Reviews, Volume 2, Issue 4, December 2005Dorigo M., Stutzle T., Ant Colony Optimization, Ant Colony Optimization, MIT Press 2004

Marco Dorigo (1992). Optimization, Learning and Natural Algorithms. Ph.D.Thesis, Politecnico di Milano, Italy, in Italian.The Metaphor of the Ant Colony and its Application to Combinatorial Optimization

Based on theoretical biology work of Jean-Louis Deneubourg (1987) From individual to collective behavior in social insects . Birkhuser Verlag, Boston.

Picture from http://www.scs.carleton.ca/~arpwhite/courses/95590Y/notes/SI%20Lecture%203.pdf

Applications Bus routes, garbage collection, delivery routes Machine scheduling: Minimization of transport time for

distant production locations Feeding of lacquering machines Protein folding Telecommunication networks: Online optimization Personnel placement in airline companies Composition of products

Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26lecture10_ACO.pdfSlide 1Folie 5Folie 6Stigmergy in HumansFolie 8Folie 9Folie 10Folie 25Folie 11Folie 12Folie 13Folie 14ApplicationsFolie 16Slide 15Folie 18Slide 17Folie 20Folie 21Folie 22Folie 23Folie 24Folie 26Folie 27Folie 28Folie 29Some general considerationsVehicle routing problemsSlide 29Folie 3Negative pheromones in real antsSlide 32