Ant colony algorithm

18
Prepared by: Qusai Nsour CS master – Artificial inelegant course. Supervised by Dr. Khalaf Khatatneh. Summer semester- 2016

Transcript of Ant colony algorithm

Page 1: Ant colony algorithm

Prepared by:Qusai Nsour

CS master – Artificial inelegant course.Supervised by Dr. Khalaf Khatatneh.

Summer semester- 2016

Page 2: Ant colony algorithm

تعالى : قال

Page 3: Ant colony algorithm

Definition

• Ant colony algorithm is an algorithm that depends on analyzing the way ants move to find there food, and conclude a mathematical form that could be applied in solving problems that mimic the ants when travelling between point A to B or “ 2 nodes”.

Page 4: Ant colony algorithm

Why ants?

• A deep look into the way ants move in finding there way indicates that there is a systematic – not random – way when they move between there colony and food ( 2 nodes).

• Ants tend to use the shortest path between the two nodes.

• If there happened to be an obstacle in there way, ants learn how to avoid it.

Page 5: Ant colony algorithm

Historical look

Page 6: Ant colony algorithm

The ants trick

• In fact there is no trick behind this, the ants depends on there smelling sense, they follow the smell that is left on the way to the food by a leading “agent”.

• This smell is generated by a “pheromone” :A pheromone is a unseen or excreted

chemical factor that triggers a social response in members of the same species.

Page 7: Ant colony algorithm

Organizing randomness

• Ants start searching for food in a random way, there is no clue to where the food is.

Page 8: Ant colony algorithm

Organizing randomness

There graph is generated in a random way“what does this look like?!!”colony

food

Page 9: Ant colony algorithm

Organizing randomness

Eliminating NodesBy finding where the smell of the pheromone is

stronger

Page 10: Ant colony algorithm

Organizing randomness

Determining shortest path

Page 11: Ant colony algorithm

Here comes the AI …!

• We have a multi-agent system• A learning techniques• Factors that define those techniques • We have source and destination or shall we

say “a game”…!• Can we build a computational model for this

system … if yes … congratulations … this means that we can program it…!!!

Page 12: Ant colony algorithm

The math

Page 13: Ant colony algorithm

The math

Page 14: Ant colony algorithm

Main ant colony alg.

Page 15: Ant colony algorithm

Applications

Page 17: Ant colony algorithm

refrences

• https://en.wikipedia.org/wiki/Main_Page• Dorigo, Marco, and Luca Maria Gambardella.

"Ant colony system: a cooperative learning approach to the traveling salesman problem."Evolutionary Computation, IEEE Transactions on 1.1 (1997): 53-66.

• Anirudeh, et, Artificial intelligent seminar, Indian institute of technology.

Page 18: Ant colony algorithm

• Any Questions ?