Ants are able to find the shortest path between their nest and a food source based on pheromone trails. In experiments, ants were given a choice between two paths of different lengths; they would initially choose randomly but then nearly all traffic would converge on the shorter path due to more frequent pheromone laying. Even when paths were opened at different times, ants preferred the path with more established pheromone trails. This self-organized behavior inspired the development of ant colony optimization algorithms for solving complex optimization problems.

1. HowDo Ants Find the Shortest Path?
SURESHR. JAMBAGI
Ph.D.ResearchScholar
Dept.of Entomology
UAS,GKVK-Bangalore-65
ICAR-NBAIR,Bangalore
Email: jambagisuru@gmail,com

2. Ants…??
 Most familiar social insects
 Everywhere – from the arctic tree line to the tip of South Africa
and Tasmania.
Why familiar- walk about on the ground in disciplined moving
columns
 Collective foraging
So far, 13,000 species of ants- catalogued, named and described.
 Ants drop pheromone while travelling, and make distributed
decisions based only on indirect interaction and local information.
2

3. Myrmecologists
3

4. Need of shortcut route…!!!
 Finding the shortest route is extremely important not only for Roman road
builders, thirsty rugby men and applied mathematicians working on this
very problem, but also for any animal that must move regularly between
different points.
 How can an animal with only limited and local navigational information
achieve this?
 Jean-Louis Deneubourg and his colleagues- How ants find the shortest
path to food
4

5. Jean-Louis Deneubourg
 Theoretical chemist but has always been
interested in animal behaviour.
 PhD in the Université de Bruxelles in Belgium
 Student of- Romanovich Prigogine (physical
chemist & Nobel laureate)
 Wanted to work on ants!
 Teammates: J M Pasteels, S Aron and S Goss (3-
brilliant biologists)
His words:
Deneubourg says he was
interested in shortcuts because
he thought “time is money”,
5

6.  He was more intrigued by the columns of moving ants rather than their diversity.
 So, he decided to work on “How Ants Find the Shortest Path?”
Deneubourg et al. decided to conduct laboratory experiments to understand how these
ants choose the shorter of two available paths.
 Ant: Argentinian ant- Iridomyrmex humilis (Linepithema humile)
 Why- Notoriously invasive in Europe and other places.
6

7. Experimental set-up
Model reconstructed by UG students of Dr. Gadagkar at IISc
7

8.  Results:
“Ants began by randomly using either branch but soon, nearly all
the traffic was on the shorter branch”.
o How did the ants manage to achieve this feat????
- Biologist answer: Ants may measure and compare distances and that
natural selection would have selected for ants that choose the shortest
path. (TRUE….!!!??)
YES…..Ants and bees can indeed measure distances BUT not the
justification for above
8

9.  But being a chemist, Deneubourg appears to have followed his
natural instinct to pursue a physicochemical explanation, rather than a
biological one.
WHAT WOULD IT BE…??
 Well known fact:
“Many ant species including the Argentine ants deposit volatile
chemicals, known as ‘pheromones’ as they walk from their nest to the
food and from the food to the nest”.
9

10. CASE-I
Fig. 1: Argentine ants are given a bridge with two equal branches
10

11. Outcomes…?
 Most of the traffic was restricted to one of the two branches (either the right or the left) in 92% of the
experiments while the traffic was distributed roughly equally between the two branches only in 8% of
the experiments.
WHY…?
 Branch with the slightly higher concentration of pheromone will be slightly more attractive to the ants:
Leading to slightly more traffic on that branch.
This will result in even higher pheromone concentration on that branch and the process will runaway due
to positive feedback resulting in that branch winning completely over its counterpart
Completely Random
11

12. CASE-II
Fig. 2: Argentine ants are given a bridge with two unequal branches
12

13. Outcomes…?
 Ants choosen one of the two branches randomly in the beginning.
 Later, more traffic per unit time on the shorter branch
WHY…?
 Ants choosing the shorter branch will reach the food sooner and return sooner and thus make
more trips per unit time.
 Here, there will be more pheromone build up on the shorter branch and through a similar
positive feedback, the shorter branch will win over the longer branch.
ANT RULE: Lay pheromone and follow the trails of others
13

14. CASE-III
Fig. 3: Argentine ants are given a bridge with two unequal branches, but the short
branch is opened 30 minutes after the ants have been using the long branch. 14

15. Outcomes…?
 Short branch fails to win over long branch.
There was less than 50% of the traffic on the short path in more than 79% of the
experiments (sum of the two left-most bars).
WHY…?
 So much pheromone would have been built upon the long branch by the time the
short branch was opened, that it would not be able to win over the long branch.
SELF ORGANISATION
15

16. How Do Ants Find the Shortest Path?
 How ants choose the shortest path is the kind of curiosity-based question that any person
should be able to ask; it does not take training of any kind.
 Instead, based on the relative quantity of pheromone deposit, ants can able to follow the
shortest path available.
ANSWER
16

17. Useless Knowledge Can Have Unexpected Uses…??
Marco Dorigo: (1961- )
 Now Director of IRIDIA, the Artificial Research Intelligence
Laboratory in the University of Brussels.
 He was a young student pursuing his PhD when Deneubourg
and his colleagues were performing these experiments.
 Applied this knowledge in developing efficient computer
algorithms for solving many discrete optimization problems.
17

18.  Dorigo and colleagues considered the problem of how artificial ants can
move around on graphs of varying complexity, and find the shortest path
between any two points.
 Thus, they first developed an algorithm they called a ‘simple ant colony
optimization’ or S-ACO: used to understand in detail, how ant colony
optimization works.
 Computer algorithm: ANT COLONY OPTIMISATION (ACO) –
metaheuristic
18

19. Application of ACO
 Internet and telecommunication
 Solve travelling salesman problem
 Adaptive routing in communication networks
 Distributed algorithms for data clustering
 Dynamic resource sharing
 Machine scheduling
Vehicle routing
Sequence learning
Machine learning
Graph colouring
19

20. REFERENCES
Gadagkar, R., 2021. Experiments in animal behaviour: Cutting-Edge
Research at Trifling Cost. Indian Academy of Sciences,
Bengaluru. pp. 29-42.
Shah, S., Kothari, R. and Chandra, S., 2011. Debugging ants: How ants
find the shortest route. In 2011 8th International Conference on
Information, Communications & Signal Processing (pp. 1-5).
IEEE.
20

21. 21
TIME TO END
THANK YOU ALL...