meenu_sharma_swarmint_19.03.10

34
SWARM INTELLIGENCE SWARM INTELLIGENCE From Natural to Artificial Systems

Upload: meenusharma06

Post on 06-Apr-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 1/34

SWARM INTELLIGENCESWARM INTELLIGENCEFrom Natural to Artificial Systems

Page 2: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 2/34

OVER VIEWOVER VIEWy Introduction

y History

y Algorithms

y Taxonomyy Characteristics

y Applications

y

Scopey Limitations

y Recap

Page 3: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 3/34

INTR ODUCTIONINTR ODUCTION

y Swarm intelligence (SI) describes the

collective behaviour of decentralized, self-

organized systems, natural or artificial.

y The concept is employed in work on

artificial intelligence.

Page 4: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 4/34

SISI -- DEFINITIONDEFINITION

³any attempt to design algorithms or distributed   problem-solving devices inspired by the collective 

behaviour of social insect colonies and other 

animal societies ́ [Bonabeau, Dorigo,

Theraulaz: Swarm Intelligence]

Page 5: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 5/34

BRIEF HISTORYOF SIBRIEF HISTORYOF SI

y One of the first researchers in this field

was French biologist Grassé.

y In 1970s ARPA financed research projects

on the first multi-agent systems with the

Hearsay-II blackboard project.

Page 6: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 6/34

HISTORY CONTD«.HISTORY CONTD«.

y In this decade the ³Actor Model´ was

invented by Carl Hewitt, Peter Bishop and

Richard Steiger.

y The Multi-Agent Systems Lab at Amherst

held the first workshop on ³distributed

artificial intelligence´ in 1980.

Page 7: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 7/34

SWARMINGSWARMING ±  ± EXAMPLEEXAMPLE

y Bird Flocking.

y ³Boids

´model was proposed by Reynolds

Boids = Bird-oids (bird like)

y Only three simple rules .

Page 8: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 8/34

COLLISION

AVOIDANCE

FLOCK CENTERING

VELOCITY

MATCHING

Page 9: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 9/34

Collision AvoidanceCollision Avoidance

y Rule 1: Avoid Collision with neighbouring

 birds .

Page 10: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 10/34

Flock CenteringFlock Centering

y Rule 2: Stay near neighbouring birds.

Page 11: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 11/34

Velocity MatchingVelocity Matching

y Rule 3: Match the velocity of 

neighbouring birds.

Page 12: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 12/34

EXAMPLE ALGORITHMSEXAMPLE ALGORITHMS

1. Ant Colony O ptimization.

2. Particle Swarm O ptimization.3. Stochastic Diffusion Search.

4. Gravitational Search Algorithm.

Page 13: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 13/34

1.1. Ant Colony OptimizationAnt Colony Optimization --

Biological InspirationBiological Inspiration

y Inspired by foraging behaviour of ants.

y Ants find shortest path to food source

from nest.

y Ants deposit pheromone along travelled

 path.

y Has adaptability,robustness and

redundancy.

Page 14: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 14/34

Foraging behavior of AntsForaging behavior of Ants

2 ants start with equal probability of going

on either path.

Page 15: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 15/34

Foraging behavior of AntsForaging behavior of Ants

The ant on shorter path has a shorter to-

and-fro time from it¶s nest to the food.

Page 16: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 16/34

Foraging behavior of AntsForaging behavior of Ants

The density of pheromone on the shorter path is

higher because of 2 passes by the ant (as compared to

1 by the other).

Page 17: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 17/34

Foraging behavior of AntsForaging behavior of Ants

The next ant takes the shorter route.

Page 18: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 18/34

Foraging behavior of AntsForaging behavior of Ants

Over many iterations, more ants begin using the path with higher pheromone, thereby further 

reinforcing it.

Page 19: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 19/34

Foraging behavior of AntsForaging behavior of Ants

After some time, the shorter path is almost

exclusively used.

Page 20: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 20/34

Ant Colony MetaheuristicAnt Colony Metaheuristic

Page 21: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 21/34

2. Particle Swarm Optimization2. Particle Swarm Optimization

Bird flocking is one of the best example of 

PSO in nature.One motive of the development of PSO was

to model human social behavior.

Page 22: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 22/34

Particle SwarmOptimizationParticle SwarmOptimization

y In particle swarm optimization (PSO), a

set of software agents called particle

search for good solutions.

y Each particle uses its own experience and

the experience of neighbour particles tochoose how to move in the search space.

Page 23: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 23/34

3.Stochastic diffusion search3.Stochastic diffusion search

y This type of swarm intelligence is based

on the tendem-calling mechanism used by

a variety of ants.

y Individual agents update their own

  preferences while randomly testing new

hypotheses.

y This process culminates in a collective or sub-collective choice of the optimal

solution.

Page 24: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 24/34

4. Gravitational Search Algorithm4. Gravitational Search Algorithm

y GSA is constructed based on the law of 

Gravity and the notion of mass

interactions.

y The GSA algorithm uses the theory of 

 Newtonian physics and its searcher agentsare the collection of masses.

Page 25: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 25/34

TAXONOMYOF SITAXONOMYOF SI

y Natural vs. Artificial.

yScientific vs. Engineering.

Page 26: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 26/34

SWARMINGSWARMING

CHARACTERISTICSCHARACTERISTICS

y Simple rules for each individual.

y No central control.

Decentralized and hence robust.

y Emergent.

Performs complex functions.

Page 27: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 27/34

APPLICATIONS OF SIAPPLICATIONS OF SI

y Crowd simulation.

y Ant-based routing.

y

Swarm Robotics.y creation of the video sequence.

Page 28: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 28/34

SCOPE OF SISCOPE OF SI

y The U.S. military is investigating swarm

techniques for controlling unmanned

vehicles.

y The European Space Agency is thinking

about an orbital swarm for self assemblyand interferometry.

Page 29: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 29/34

SCOPE CONTD«.SCOPE CONTD«.

y A 1992 paper by M. Anthony Lewis and

George A. Bekey discusses the possibility

of using swarm intelligence to control

nanobots within the body for the purpose

of killing cancer tumors.

y NASA is investigating the use of swarmtechnology for planetary mapping.

Page 30: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 30/34

SISI -- LIMITATIONSLIMITATIONS

y Theoretical analysis is difficult, due to

sequences of probabilistic choices.

y Most of the researches are experimental.

y Though convergence is guaranteed, time

to convergence is uncertain.

Page 31: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 31/34

RECAPRECAP

y Swarm intelligence is the discipline that

deals with natural and artificial systems

composed of many individuals that

coordinate using decentralized control andself-organization.

y Provide heuristic to solve difficult

 problems .y Has been applied to wide variety of 

applications.

Page 32: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 32/34

REFERENCESREFERENCESy Reynolds, C. W. (1987) Flocks, Herds, and Schools: A

Distributed Behavioral Model, in Computer Graphics,21(4) (SIGGRAPH '87 Conference Proceedings) pages 25-34.

y

James Kennedy, Russell Eberhart. Particle SwarmO ptimization, IEEE Conf. on Neural networks ± 1995

y www.adaptiveview.com/articles/ ipsop1

y

M.Dorigo, M.Birattari, T.Stutzle, Ant colonyoptimization ± Artificial Ants as a computationalintelligence technique, IEEE ComputationalIntelligence Magazine 2006

Page 33: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 33/34

 .

D umb parts,

 properly connected into 

a swarm 

 yield  smart results 

Page 34: meenu_sharma_swarmint_19.03.10

8/3/2019 meenu_sharma_swarmint_19.03.10

http://slidepdf.com/reader/full/meenusharmaswarmint190310 34/34

THANK YOUTHANK YOU