You are on page 1of 18

Ant Colony Optimization

By: Aster Alemu


Outline
• Ant Colony Optimization (ACO)
• Discrete ACO
• Continuous ACO
Populations
Ant Colony Optimization (ACO)

• ACO is another algorithm inspired by nature


• It is applicable to both continuous and discrete problems.
• ACO is interchangeable with a genetic algorithm (GA)
• The continuous and discrete versions of ACO differ considerably
Introduction to the ACO: Authors
• Marco Dorigo (1992) introduced discrete ACO in his PhD thesis
• Christian Blum and Krzysztof Socha (2005) published a continuous
version of ACO
Introduction to the ACO: Origins
• The foraging behavior of ants influences both the discrete and
continuous versions of ACO
• In nature, ants initially wander randomly in their search for food.
Upon finding food, the ant returns to the colony while laying down
pheromone trails.
• The presence of pheromone increases the probability that an ant will
move into that location and continue on that trail.
• If other ants discover these paths, a strong probability exists that they
will not continue traveling at random (Instead, they will find the trail,
returning and reinforcing it in their search for food).
• Over time, the pheromone trail evaporates, reducing its attractive
strength. Therefore, as an ant spends more time traveling down the
path and back again, there is less time for the pheromones to
evaporate
• Consequently, ants march more frequently over a short path, causing
the pheromone density to become higher on shorter paths compared
to longer ones.
• Additionally, pheromone evaporation encourages exploration beyond
the initial paths. Without evaporation, the favorite paths of the first
ants tend to be extremely attractive to later ants
Figure 1: ants foraging for food with the majority of the ants following
the established path.
• Thus, when one ant finds a shorter path from the colony to a food
source, other ants are more likely to follow that path. This positive
feedback eventually leads to most of the ants following a single path.
Some ants will still forage randomly to find shorter paths. The ACO
algorithm mimics this behavior with simulated ants marching around
a solution graph that represents the problem to solve.
Introduction to the ACO :Concepts
Discrete ACO
Discrete ACO…
Mathematical model
• Example of discrete ACO
Algorithm of discrete ACO
Continuous ACO
• Mathematical model
• Example of discrete ACO
Algorithm of Continuous ACO

You might also like