14
Particle Swarm Optimization and Social Interaction Between Agents Kenneth Lee TJHSST 2008

Particle Swarm Optimization and Social Interaction Between Agents Kenneth Lee TJHSST 2008

Embed Size (px)

Citation preview

Particle Swarm Optimization and Social Interaction Between Agents

Kenneth LeeTJHSST 2008

Overview

Overview of PSO

Background Research

Project Goals

Types of Social Interactions

Project State/Results

Conclusion

Overview Of PSO

Originally designed to model birdsOvertime became more analogous to a swarming animal (bees)Search for Global OptimaInfinite search spaces

Overview Of PSO

“Particles” (vectors)Random PositionRandom VelocityInfluences on Velocity

Cognitive Influence Social Influence

Convergence(?)

Particle Swarm Optimization

for k = 1 to number of particles n doif (fitness(k) < fitness_lbest(k))

lbest(k) = pos(k)endif

end do

for k = 1 to number of particles n dosocial(k)

enddo

for k = 1 to number of particles n dofor I = 1 to number of dimensions d do

R1 = randomNumberR2 = randomNumberV[k][I] = w * (C1 * R1 *

(pos-lbest) + C2 * R2 * (pos-gbest))X[k][I] = pos + V[k][I]

enddoenddo

Determining lbest

Social Interaction

Adjusting Position

Importance of Social Interaction

Influences Velocity V = ??? X’ = X + V

Encourages ExplorationThrough Social Interaction, Particles are able to communicate information and extrapolate data about the objective function.

Social Interactions

Variance of k value (# of neighbors)

Through research k values between 3-5 seem to work best

Topology? Cliques Random Share/Follow

Project State

5 Interactions NIPS SIPS RIPS FIPS DIPS

3 Benchmark Functions

Rastrigin, Six Camel Hump, Sphere

NIPS (Non-Informed Particle Swarm)

0 5 10 15 20 25

0

50

100

150

200

250

RN-3-4--1_0.txt

RN-3-4--1_1.txt

RN-3-4--1_2.txt

RN-3-4--1_3.txt

RN-3-4--1_4.txt

SIPS(Singly-Informed Particle Swarm)

0 20 40 60 80 100 120 140 160

0

50

100

150

200

250

RS-3-4--1_0.txt

RS-3-4--1_1.txt

RS-3-4--1_2.txt

RS-3-4--1_3.txt

RS-3-4--1_4.txt

RIPS(Ring Informed Particle Swarm)

0 100 200 300 400 500 600 700 800

0

50

100

150

200

250

300

350

400

RR-3-4--1_0.txt

RR-3-4--1_1.txt

RR-3-4--1_2.txt

RR-3-4--1_3.txt

RR-3-4--1_4.txt

FIPS(Fully Informed Particle Swarm)

0 5 10 15 20 25 30 35 40 45 50

0

50

100

150

200

250

300

RF-3-4--1_0.txt

RF-3-4--1_1.txt

RF-3-4--1_2.txt

RF-3-4--1_3.txt

RF-3-4--1_4.txt

DIPS(Dynamically Informed Particle Swarm)

0 50 100 150 200 250

0

50

100

150

200

250

300

350

RD-3-4--1_0.txt

RD-3-4--1_1.txt

RD-3-4--1_2.txt

RD-3-4--1_3.txt

RD-3-4--1_4.txt

Conclusions

DIPS seems to perform best

Time onlyDIPS and RIPS have 100% success rateFIPS converges fastest