Ant colony optimization
From DDL Wiki
(→External Links) |
(→Real World Ant Behavior) |
||
| Line 5: | Line 5: | ||
==Real World Ant Behavior== | ==Real World Ant Behavior== | ||
| + | When searching for food, ants will wander around randomly until they find a food source, and then return to the colony while laying down a pheromone path that can be retraced. The ants at the colony that are wandering around will pick up this trail and follow it to the food and if the food is found they will also reinforce the trail by laying down a pheromone path. | ||
| + | |||
| + | The pheromone path is not permanent and evaporates over time. This means that the paths that are longer will be weaker since there is more time between pheromone dropping. However, a short path will have a much stronger pheromone path since it is closer and will be replenished with pheromones continuously. Thus the shorter paths will be preferred and the ants will tend to travel to the optimum food source. | ||
==Application== | ==Application== | ||
Revision as of 12:44, 29 November 2006
Ant Colony Optimization (ACO) is a probabalistic technique inspired by the way ants make and find paths from the colony to food. The technique is used to solve discrete optimization problems that can be reduced to finding good paths through graphs.
Contents |
History
The first appearance of an ACO system was in a Ph.D. thesis in 1992 by Marco Dorigo at Politecnico di Milano. It was called Ant System (AS). Since 1995 various other extended versions of AS have been developed, including Ant Colony System (ACS) and MAX-MIN Ant System (MMAS). In 1999 Dorigo proposed the Ant Colony Optimization metaheuristic this is most common today.
Real World Ant Behavior
When searching for food, ants will wander around randomly until they find a food source, and then return to the colony while laying down a pheromone path that can be retraced. The ants at the colony that are wandering around will pick up this trail and follow it to the food and if the food is found they will also reinforce the trail by laying down a pheromone path.
The pheromone path is not permanent and evaporates over time. This means that the paths that are longer will be weaker since there is more time between pheromone dropping. However, a short path will have a much stronger pheromone path since it is closer and will be replenished with pheromones continuously. Thus the shorter paths will be preferred and the ants will tend to travel to the optimum food source.
Application
External Links
- Ant Colony Optimization Home Page
- ANSI C implementation of ACO, by Thomas Stützle
- Java implementation of ACO, Andrew Carson
- A list of dissertations relevant to the ACO field
- Combinatorial Optimisation - JAVA Applet demonstrating ACO variants including Ant Systems, Rank Based Ant Systems, Max-Min Ant Systems, Ant Colony Systems
References
- Éric Bonabeau, Marco Dorigo et Guy Theraulaz, Inspiration for optimization from social insect behaviour, Nature, 406 39-42, July 2000.
- F. Moyson, B. Manderick, The collective behaviour of Ants : an Example of Self-Organization in Massive Parallelism, Proceedings of AAAI Spring Symposium on Parallel Models of Intelligence, Stanford, California, 1988.
- A. Colorni, M. Dorigo et V. Maniezzo, Distributed Optimization by Ant Colonies, Proceedings of the first european conference on artificial life, Paris, France, Elsevier Publishing, 134-142, 1991.
- M. Dorigo, Optimization, Learning and Natural Algorithms, PhD thesis, Politecnico di Milano, Italy, 1992.
- Éric Bonabeau, Marco Dorigo et Guy Theraulaz, Swarm Intelligence: From Natural to Artificial Systems, Oxford University Press, 1999. ISBN 0-19-513159-2
- M. Dorigo, T. Stützle, Ant Colony Optimization, MIT Press, 2004. (ISBN 0-262-04219-3)
