You are on page 1of 34

Particle Swarm Optimization (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 (socialpsychologist) 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 adjusts its flying according to its own flying experience as well as the flying experience of other particles.

Particle Swarm Optimization (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 Swarm Optimization (PSO)


y vk vk+1 sk sk+1 vgbest vpbest

x Fig.1 Concept of modification of a searching point by PSO sk : current searching point. sk+1: modified searching point. vk: current velocity. vk+1: modified

Particle Swarm Optimization (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 particles position can be mathematically modeled according the following equation :

Vik+1 = wVik +c1 rand1() x (pbesti-sik) + c2 rand2() x (gbest-sik) .. (1)


where, vik : velocity of agent i at iteration k, w: weighting function, cj : weighting factor, rand : uniformly distributed random number sik : current position of agent i at pbesti : pbest of agent i,

between 0 and 1, iteration k,

Particle Swarm Optimization (PSO)


The following weighting function is usually utilized in (1) w = wMax-[(wMax-wMin) x iter]/maxIter where wMax= initial weight, wMin = final weight, maxIter = maximum iteration number, iter = current iteration number. sik+1 = sik + Vik+1 (3) (2)

Comments on the Inertial weight factor:


A large inertia weight (w) facilitates a global search while a small inertia weight facilitates a local search. By linearly decreasing the inertia weight from a relatively large value to a small value through the course of the PSO run gives the best PSO performance compared with fixed inertia weight settings.

Particle Swarm Optimization (PSO)

Larger w ----------- greater global search ability Smaller w ------------ greater local search ability.

Particle Swarm Optimization (PSO)


Flow chart depicting the General PSO Algorithm:
Start Initialize particles with random position and velocity vectors. Loop until all particles exhaust For each particles 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 max iter

Comparison with other evolutionary computation 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 of 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 LICP 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 Suns Optimized parameters. )

c1=c2=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 LICP 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


Layer 1 2 3 4 Pitch Period Height (inches) (inches) (inches) a b h
0.34493 60 0.37984 69 0.37984 69 0.34493 60
Dielectric constants:

Line Width (inches) w1 w2


2.25013 51E-02 5.39996 61E-03 5.39996 61E-03 2.25013 51E-02 1.59840 01E-02 9.42966 53E-03 9.42966 53E-03 1.59840 01E-02

0.72833 82 0.87043 98 0.87043 98 0.72833 82

0.25205 65 0.47074 20 0.47074 20 0.25205 65

dielectr Spacer ic (inches) (inches) loi li


8.47050 18E-03 3.46763 25E-03 3.46763 25E-03 8.47050 18E-03 0.54281 28 0.42118 45 0.54281 28

----

i = 2.55 0 i = 1.15

4 Layers for CP
Metal Layer Dielectric Sheet Spacer

Application of PSO ALGORITHM to Optimize a Meander-line Polarizer for LICP 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


Layer 1 2 3 4 Pitch Period Height (inches) (inches) (inches) a b h
0.25289 13 0.54257 24 0.54257 24 0.25289 13
Dielectric constants:

Line Width (inches) w1 w2


4.04561 83E-02 5.03585 79E-02 5.03585 79E-02 4.04561 83E-02 2.77042 02E-02 4.49816 77E-02 4.49816 77E-02 2.77042 02E-02

0.80816 69 0.95296 58 0.95296 58 0.80816 69

0.24801 43 0.34337 22 0.34337 22 0.24801 43

dielectr Spacer ic (inches) (inches) loi li


4.11472 52E-03 3.95937 12E-03 3.95937 12E-03 4.11472 52E-03 0.45338 23 0.44821 95 0.45338 23

----

i = 2.55 0 i = 1.15

4 Layers for CP
Metal Layer Dielectric Sheet Spacer

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 Suns Optimized parameters.)

c1=c2=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


Layer 1, 5 2, 6 3, 7 4, 8 Pitch Period Height (inches) (inches) (inches) a b h
0.36318 78 0.38481 70 0.38481 70 0.36318 78 1.02060 0 0.82257 65 0.82257 65 1.02060 0 0.21480 45 0.43749 99 0.43749 99 0.21480 45

Line Width (inches) w1 w2


2.86068 38E-02 4.15424 74E-02 4.15424 74E-02 2.86068 38E-02 3.07097 08E-02 3.49429 29E-02 3.49429 29E-02 3.07097 08E-02

dielectr Spacer ic (inches) (inches) loi l i


2.16235 70E-02 4.06607 84E-02 4.06607 84E-02 2.16235 70E-02 0.45507 97 0.39594 68 0.45507 97

----

Dielectric constants:

i = 2.55 0 i = 1.15

8 Layers for LP
Metal Layer Dielectric Sheet Spacer

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


Layer 1, 5 2, 6 3, 7 4, 8 Pitch Period Height (inches) (inches) (inches) a b h
0.31508 69 0.40858 01 0.40858 01 0.31508 69 1.05559 6 1.16457 6 1.16457 6 1.05559 6 0.27475 69 0.35656 08 0.35656 08 0.27475 69

Line Width (inches) w1 w2


3.32023 91E-02 6.08112 51E-02 6.08112 51E-02 3.32023 91E-02 3.20107 45E-02 0.11262 98 0.11262 98 3.20107 45E-02

dielectr Spacer ic (inches) (inches) loi l i


2.22992 49E-02 3.11277 17E-02 3.11277 17E-02 2.22992 49E-02 0.35447 25 0.31593 89 0.35447 25

----

Dielectric constants:

i = 2.55 0 i = 1.15

8 Layers for LP
Metal Layer Dielectric Sheet Spacer

END

You might also like