Kavita Singh

CS-A
0509710047
What is Swarm Intelligence
(SI)?

“The emergent collective

intelligence of groups of
simple agents.”
 group foraging of social insects
 cooperative transportation
 division of labour
 nest-building of social insects
 collective sorting and clustering
The computer revolution changed human
societies:
 Communication

 Transportation

 Industrial production

 Administration, writing and bookkeeping

 Technological advances

 Entertainment

However, some problems cannot be tackled

with traditional hardware and software!
Computing tasks have to be :-
 Well-defined
 Fairly predictable
 Computable in reasonable time with serial

computers.
Well-defined, but computational hard
problems
NP hard problems (Travelling Salesman
Problem)
Action-response planning (Chess playing)
 DNA based computing (chemical
computation) is inspired by the
human evolution.
 Artificial neural network is a

simplified model of human brain.

 Bio-computing(simulation of

biological mechanisms).
1. Self-Organization is based on:-

 Positive feedback(amplification)
 Negative feedback (for balancing)
 Amplification of fluctuations(random walks,
errors)
 Multiple interactions

2. Stigmergy:-
Indirect communication via interaction with
environment.
 Species lay chemical substance pheromone
while travelling from nest, to nest or
possibly in both directions.

 Pheromones evaporate.

 Pheromones accumulate with multiple ants

using same path.
The natural behavior of these ants and be programmed into an
ant algorithm, which we can use to find the shortest path within
graphs.
As ants move they leave behind a chemical substance called
pheromone, which other ants can smell and identify that an ant
has been there before.
 distributed system of interacting
autonomous agents
 goals: performance optimization and

robustness
 self-organized control and cooperation

(decentralized)
 division of labor and distributed task

allocation
 indirect interactions
It is a 3-step process.

 Identification of analogies: in swarm

biology and IT systems.
 Understanding: computer modeling of

realistic swarm biology.

 Engineering: model simplification and

tuning for IT applications.

 Complex NP complete problems.
 Vehicle routing.
 Network maintenance.
 The traveling salesperson.
 Computing the shortest route

between two points.

 Learning algorithms developed with
artificial intelligence systems such as neural
networks but imperfections and
inefficiencies in both the hardware and
software have prevent reliable results.
 Genetic algorithms also made an attempt at

these problems, and had some success.

The algorithms were considered too
complex to re-implement.
1. Visit cities in order to make
sales.
2. Save on travel costs.
3. Visit each city once
(Hamiltonian circuit).
 If there are N cities, then the number of
different paths among them is 1.2……(N-1).

 Time to examine single path = N.

 Total time to perform the search = (N-1)!

 For 10 cities, time reqd. = 10! = 3,268,800.

 Use of agents for TSP problem.
 They sense and dispense pheromone.
 Memory to back step through the
graph.
 Each agent starts at a random starting
city.
 Once agent finishes a tour, it
determines the size of the tour.
 Then pheromone is added to the tour,
the shorter the tour, the higher the
pheromone level.
 No guarantee that the first tour the
agents will converge the shortest
path.
 Agents explore other tours.
The “stray” agent finds a shorter
path.
Adjusts the pheromone levels.
Plenty of computing time needed to
converge on the optimal tour.
 The ant algorithm approach will still
solve faster than other algorithms.
 Vehicle routing is similar to the
TSP problem.
 Employee services the client by

going to them.
 Minimize cost.
 Use the same optimal

Hamiltonian circuit as in the TSP

problem.
 The economy is an example of SI
that most researchers forget to
consider.
 SI demonstrates complex behavior

that arises from simple individual

interactions.
 No one can control the economy,

as there are no groups that can

consistently control the economy.
 The reaction of the population causes the
the economy to slow down.

 Simulating an economy using ant

algorithms.

 Makes it possible to control or predict the

ebb and flow of this complex behavior.

 Swarm intelligence, is still in its infancy.

 A project such as simulating the economy is

still far beyond the its capability.
 Scientists are realizing SI’s potential.

 The use of ant algorithms within computing

systems has helped to solidify swarm
intelligence’s place in the computing world.

 Already researchers are observing other

social animals, such as bees and schools of
fish in order to utilize it in future
applications and algorithms.
 http://www.engr.iupui.edu/~eberhart/

 http://users.erols.com/cathyk/jimk.html

 http://www.alife.org/

 http://www.aridolan.com/

 http://www.red3d.com/cwr/boids/

 http://iridia.ulb.ac.be/~mdorigo/ACO/ACO.html

 http://www.engr.iupui.edu/~shi/Coference/psopap4.html