You are on page 1of 34
Particle
Particle Swarm
Swarm Optimization
Optimization (PSO)
(PSO)
PSO is a robust stochastic optimization technique based on
the movement and intelligence of swarms.
PSO applies the concept of social interaction to problem
solving.
It was developed in 1995 by James Kennedy (social-
psychologist) and Russell Eberhart (electrical engineer).
It
uses
a number of agents (particles) that constitute a
swarm moving around in the search space looking for the
best solution.
Each particle is treated as a point in a N-dimensional space
which
its
“flying”
according
to
its
own flying
experience as well
particles.
as
the
flying
experience
of
other
Particle
Particle Swarm
Swarm Optimization
Optimization (PSO)
(PSO)
• Each particle keeps track of its coordinates in the solution
space which are associated with the best solution (fitness)
that has achieved so far by that particle. This value is called
personal best , pbest.
Another best value that is tracked by the PSO is the best
value obtained so far by any particle in the neighborhood of
that particle. This value is called gbest.
The basic concept of PSO lies in accelerating each particle
toward its pbest and the gbest locations, with a random
weighted accelaration at each time step as shown in Fig.1
Particle
Particle Swarm
Swarm Optimization
Optimization (PSO)
(PSO)
y
s k+1
s
k+1
v
v
k
k
v
v
k+1
k+1
v
v
gbest
gbest
v
v
pbest
pbest
s s k
k
x
Fig.1 Concept of modification of a searching point by PSO
s k : current searching point.
s k+1 : modified
searching point.
v
k : current velocity.
v k+1 : modified
Particle
Particle Swarm
Swarm Optimization
Optimization (PSO)
(PSO)
• Each particle tries to modify its position using the following
information:
 the current positions,
 the current velocities,
 the distance between the current position and pbest,
 the distance between the current position and the gbest.
• The modification of the particle’s position can be mathematically
modeled according the following equation :
V i k+1 = wV i k +c 1 rand 1 (…) x (pbest i -s i k ) + c 2 rand 2 (…) x (gbest-s i k ) …
..
(1)
where,
v i k : velocity of agent i at iteration k,
w: weighting function,
between 0 and 1,
iteration k,
c j : weighting factor,
rand : uniformly distributed random number
s i k : current position of agent i at
pbest i : pbest of agent i,
gbest: gbest of the
Particle
Particle Swarm
Swarm Optimization
Optimization (PSO)
(PSO)
The following weighting function is usually utilized in (1)
w = wMax-[(wMax-wMin) x iter]/maxIter
(2)
where
wMax= initial weight,
wMin = final weight,
maxIter = maximum iteration number,
iter = current iteration number.
s i k+1
=
s i k +
V i k+1
(3)
Particle
Particle Swarm
Swarm Optimization
Optimization
(PSO)
(PSO)
the Inertial
Inertial weight
weight factor:
factor:
AA large
large inertia
inertia weight
weight ((ww)) facilitates
facilitates aa global
global search
search while
while
aa small
small inertia
inertia weight
weight facilitates
facilitates aa local
local search.
search.
ByBy linearly
linearly decreasing
decreasing the
the inertia
inertia weight
weight from
from aa relatively
relatively
large
large value
value toto aa small
small value
value through
through the
the course
course of
of the
the
PSO
PSO run
run gives
gives the
the best
best PSO
PSO performance
performance compared
compared
with
with fixed
fixed inertia
inertia weight
weight settings.
settings.
Larger
Larger ww -----------
----------- greater
greater global
global search
search ability
ability
Smaller
Smaller ww ------------
------------ greater
greater local
local search
search ability.
ability.
Particle
Particle Swarm
Swarm Optimization
Optimization (PSO)
(PSO)
Flow chart depicting the General PSO Algorithm:
Start
Initialize particles with random position
and velocity vectors.
For each particle’s position (p)
evaluate fitness
If fitness(p) better than
fitness(pbest) then pbest= p
Set best of pBests as gBest
Update particles velocity (eq. 1) and
position (eq. 3)
Stop: giving gBest, optimal solution.
Loop until all
particles exhaust
Loop until max iter
Comparison
Comparison with
with other
other evolutionary
evolutionary
computation
computation techniques.
techniques.
Unlike in genetic algorithms, evolutionary programming and
evolutionary strategies, in PSO, there is no selection operation.
All particles in PSO are kept as members of the population through
the course of the run
PSO is the only algorithm that does not implement the survival of
the fittest.
No crossover operation in PSO.
eq 1(b) resembles mutation in EP.
In EP balance between the global and local search can be adjusted
through the strategy parameter while in PSO the balance is
achieved through the inertial weight factor (w) of eq. 1(a)
Variants
Variants of
of PSO
PSO
Discrete PSO ……………… can handle discrete binary
variables
MINLP PSO………… can handle both discrete binary and
continuous variables.
• Hybrid PSO…………. Utilizes basic mechanism of PSO
and the natural selection mechanism, which is usually
utilized by EC methods such as GAs.
Application of PSO ALGORITHM to Optimize a
Meander-line Polarizer for LI→CP conversion
Intialization parameters used for PSO:
wMax=0.41
wMin=0.4
(Note:The inertial weight ,w is linearly decreased from wMax to wMin
according the Eq. (2), w is chosen virtually constant in this case for better
local search near the Sun’s Optimized parameters. )
c
1 =c 2 =1.49
maxIter=2000
The above parameters are used in conjuction with eqs.
(1) & (2)
Swarm size/Population size used for solution search : 25
Application of PSO ALGORITHM to Optimize a
Meander-line Polarizer for LI→CP conversion
Frequency band of interest: 3.5 to 6.5 (GHz)
(evaluated at 12 frequency points)
Desired VSWR <= 1.2
Desired AR <= 0.5 (dB)
Total number of fitness evaluations: 100025
Note: For my implementation of the PSO the number of fiteness evaluations are
calculated as follows: (2 x swarmsize x maxIter)+ swarmsize = (2 x 25 x 2000)+ 25
The following slides include the results for the broadband case.
Mean best & Best fitness over 50 runs
VSWR
Axial Ratio (dB)
Optimized dimensions for 4-layer
Meander Line Polarizer
Pitch
Pitch
Period
Period
Height
Height
Line
Line Width
Width
Spacer
Spacer
dielectric
dielectric
Layer
Layer
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
aa
bb
hh
ll ii
ww 11
ww 22
ll oi oi
2.2501351E-
2.2501351E-
1.5984001E-
1.5984001E-
8.4705018E-
8.4705018E-
11
0.3449360
0.3449360
0.7283382
0.7283382
0.2520565
0.2520565
0.5428128
0.5428128
0202
0202
0303
5.3999661E-
5.3999661E-
9.4296653E-
9.4296653E-
3.4676325E-
3.4676325E-
22
0.3798469
0.3798469
0.8704398
0.8704398
0.4707420
0.4707420
0.4211845
0.4211845
0303
0303
0303
5.3999661E-
5.3999661E-
9.4296653E-
9.4296653E-
3.4676325E-
3.4676325E-
33
0.3798469
0.3798469
0.8704398
0.8704398
0.4707420
0.4707420
0.5428128
0.5428128
0303
0303
0303
2.2501351E-
2.2501351E-
1.5984001E-
1.5984001E-
8.4705018E-
8.4705018E-
44
0.3449360
0.3449360
0.7283382
0.7283382
0.2520565
0.2520565
----
----
0202
0202
0303
Dielectric constants:
 2.55
 1.15
 i
0i
4 Layers for CP
Metal
Dielectric
Spacer
Layer
Sheet
Application of PSO ALGORITHM to Optimize a
Meander-line Polarizer for LI→CP conversion
Frequency bands of interest:
Band1: 3.7 to 4.2 (GHz)
Band2: 5.9 to 6.4 (GHz)
(evaluated at 2 frequency points: 3.95 (GHz), 6.15 (GHz))
Desired VSWR <= 1.2
Desired AR <= 0.5 (dB)
Total number of fitness evaluations: 100025
The following slides include the results for the dualband case.
Mean best & Best fitness over 50 runs
VSWR
Axial Ratio (dB)
Optimized dimensions for 4-layer
Meander Line Polarizer
Pitch
Pitch
Period
Period
Height
Height
Line
Line Width
Width
Spacer
Spacer
dielectric
dielectric
Layer
Layer
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
aa
bb
hh
ll ii
ww 11
ww 22
ll oi oi
4.0456183E-
4.0456183E-
2.7704202E-
2.7704202E-
4.1147252E-
4.1147252E-
11
0.2528913
0.2528913
0.8081669
0.8081669
0.2480143
0.2480143
0.4533823
0.4533823
0202
0202
0303
5.0358579E-
5.0358579E-
4.4981677E-
4.4981677E-
3.9593712E-
3.9593712E-
22
0.5425724
0.5425724
0.9529658
0.9529658
0.3433722
0.3433722
0.4482195
0.4482195
0202
0202
0303
5.0358579E-
5.0358579E-
4.4981677E-
4.4981677E-
3.9593712E-
3.9593712E-
33
0.5425724
0.5425724
0.9529658
0.9529658
0.3433722
0.3433722
0.4533823
0.4533823
0202
0202
0303
4.0456183E-
4.0456183E-
2.7704202E-
2.7704202E-
4.1147252E-
4.1147252E-
44
0.2528913
0.2528913
0.8081669
0.8081669
0.2480143
0.2480143
----
----
0202
0202
0303
Dielectric constants:
 2.55
 1.15
 i
0i
4 Layers for CP
Metal
Dielectric
Spacer
Layer
Sheet
Application of PSO ALGORITHM to Optimize a
Meander-line Polarizer for LP rotation
Intialization parameters used for PSO:
wMax=0.41
wMin=0.4
(Note:The inertial weight ,w is linearly decreased from wMax to wMin
according the Eq. (2), w is chosen virtually constant in this case for better
local search near the Sun’s Optimized parameters.)
c
1 =c 2 =1.3
maxIter=1000
The above parameters are used in conjuction with eqs.
(1) & (2)
Swarm size/Population size used for solution search : 25
Application of PSO ALGORITHM to Optimize a
Meander-line Polarizer for LP rotation
Frequency band of interest: 3.5 to 6.5 (GHz)
(evaluated at 12 frequency points)
Desired VSWR <= 1.2
Phase Difference around 180°
Total number of fitness evaluations: 50025
The following slides include the results for the broadband case.
Mean best & Best fitness over 15 runs
VSWR
Axial Ratio (dB)
Phase Difference
Optimized dimensions for 8-layer
Meander Line Polarizer
Pitch
Pitch
Period
Period
Height
Height
Line
Line Width
Width
Spacer
Spacer
dielectric
dielectric
Layer
Layer
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
aa
bb
hh
ll ii
ww 11
ww 22
ll oi oi
2.8606838E-
2.8606838E-
3.0709708E-
3.0709708E-
2.1623570E-
2.1623570E-
1,
1, 55
0.3631878
0.3631878
1.020600
1.020600
0.2148045
0.2148045
0.4550797
0.4550797
0202
0202
0202
4.1542474E-
4.1542474E-
3.4942929E-
3.4942929E-
4.0660784E-
4.0660784E-
2,
2, 66
0.3848170
0.3848170
0.8225765
0.8225765
0.4374999
0.4374999
0.3959468
0.3959468
0202
0202
0202
4.1542474E-
4.1542474E-
3.4942929E-
3.4942929E-
4.0660784E-
4.0660784E-
3,
3, 77
0.3848170
0.3848170
0.8225765
0.8225765
0.4374999
0.4374999
0.4550797
0.4550797
0202
0202
0202
2.8606838E-
2.8606838E-
3.0709708E-
3.0709708E-
2.1623570E-
2.1623570E-
4,
4, 88
0.3631878
0.3631878
1.020600
1.020600
0.2148045
0.2148045
----
----
0202
0202
0202
 1.15
Dielectric constants:
 2.55
 i
0i
8 Layers for LP
Metal
Dielectric
Spacer
Layer
Sheet
Application of PSO ALGORITHM to Optimize a
Meander-line Polarizer for LP rotation
Frequency bands of interest:
Band1: 3.7 to 4.2 (GHz)
Band2: 5.9 to 6.4 (GHz)
(evaluated at 2 frequency points: 3.95 (GHz), 6.15 (GHz))
Desired VSWR <= 1.2
Phase Difference around 180°
Total number of fitness evaluations: 50025
The following slides include the results for the dualband case.
Mean best & Best fitness over 15 runs
VSWR
Axial Ratio (dB)
Phase Difference
Optimized dimensions for 8-layer
Meander Line Polarizer
Pitch
Pitch
Period
Period
Height
Height
Line
Line Width
Width
Spacer
Spacer
dielectric
dielectric
Layer
Layer
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
(inches)
aa
bb
hh
ll ii
ww 11
ww 22
ll oi oi
3.3202391E-
3.3202391E-
3.2010745E-
3.2010745E-
2.2299249E-
2.2299249E-
1,
1, 55
0.3150869
0.3150869
1.055596
1.055596
0.2747569
0.2747569
0.3544725
0.3544725
0202
0202
0202
6.0811251E-
6.0811251E-
3.1127717E-
3.1127717E-
2,
2, 66
0.4085801
0.4085801
1.164576
1.164576
0.3565608
0.3565608
0.1126298
0.1126298
0.3159389
0.3159389
0202
0202
6.0811251E-
6.0811251E-
3.1127717E-
3.1127717E-
3,
3, 77
0.4085801
0.4085801
1.164576
1.164576
0.3565608
0.3565608
0.1126298
0.1126298
0.3544725
0.3544725
0202
0202
3.3202391E-
3.3202391E-
3.2010745E-
3.2010745E-
2.2299249E-
2.2299249E-
4,
4, 88
0.3150869
0.3150869
1.055596
1.055596
0.2747569
0.2747569
----
----
0202
0202
0202
 2.55
 1.15
Dielectric constants:
 i
0i
8 Layers for LP
Metal
Dielectric
Spacer
Layer
Sheet
END