You are on page 1of 23

EVOLUTIONARY OPTIMIZATION ALGORITHMS

By Sumanth (EC8435)

OVERVIEW
3/26/2012

What

is Optimization? How to Optimize? Deterministic Techniques Vs Evolutionary Techniques Genetic Algorithm Particle Swarm Optimization Ant Colony Optimization Differential Evolution Conclusion ( Most efficient algorithm) References

by Sumanth Kumar

WHAT IS OPTIMIZATION ?
3/26/2012

Effective selection of optimal solution. Maximum or Minimum. Cost function.


by Sumanth Kumar

Applications
Signal Processing. Communications. Computer Networks. Economics etc..

HOW TO OPTIMIZE ??
3/26/2012

Calculus tell us to differentiate and find optimum value. To differentiate cost function must be known. Numerical methods can be used, if cost function is not available. Newtons forward and backward differences can be used if sample data is known.

by Sumanth Kumar

MAJOR PROBLEM
3/26/2012

As complexity of cost function increases.

by Sumanth Kumar

Rastrigins Function

PROBLEMS WITH DETERMINISTIC TECHNIQUES


3/26/2012

Struck at local minima or maxima. Computationally complex.

global

by Sumanth Kumar

local

PROBABILISTIC OPTIMIZATION TECHNIQUES


3/26/2012

Searches N dimensional search space randomly and lands on a approximate solution. Doesnt provide ideal optimum value but gives approximate value. Evolutionary algorithms got their basic ideas from nature. Examples

by Sumanth Kumar

Survival of the fittest. Motion of birds in search of food. Ants in search of food. Bees etc..

DETERMINISTIC VS PROBABILISTIC TECHNIQUES


3/26/2012

Results when Rastrigin's function is optimized in both methods.

by Sumanth Kumar

Compared value Maximum Minimum Time (in s)

Deterministic Technique 60.3002 3.1198 74.84

Genetic Algorithm Particle Swarm Optimization 76.5207 0.0095 0.1307 72.042 1.3590 0.0564

GENETIC ALGORITHM (GA)


3/26/2012

Proposed by John Holland in 1975. Based on natural selection process Survival of the fittest. Every possible solution is called chromosome. Chromosomes are crossed over and mutated to get better offsprings. Works iteratively. Each iteration is called as Generation. Best chromosomes are passed on to next generations.

by Sumanth Kumar

GENETIC ALGORITHM (GA)

CONTD..
3/26/2012 by Sumanth Kumar

Evaluation of Individuals

10

GENETIC ALGORITHM (GA)

CONTD..
3/26/2012

Chromosome

Binary coded - 1 0 1 0 0 1 1 1 0 1 Real encoded - 1.236 2.36 4.23


Chromosome 1 Chromosome 2 Offspring 1 Offspring 2 11010011100 01100010010 11010010010 01100011100

by Sumanth Kumar

Crossover
Single point Two point

11

GENETIC ALGORITHM (GA)

CONTD..
3/26/2012

Mutation

chromosome 1 0 1 1 0 0 0 1 1 0 offspring
Convergence plot for Rastrigins Funciton

1011010110

by Sumanth Kumar

82 GA 80

Value of the cost function

78

76

74

72

70

68 0

10 15 No of iterations

20

25

12

PARTICLE SWARM OPTIMIZATION


3/26/2012

Inspired by the swarm of birds in search of food. Considers both self and group knowledge while searching. Initial population are called as particles. Each particle has a position and velocity randomly assigned. Lesser computational complexity than GA. Positions and velocities are updated in every iteration. No crossover and mutation operators.

13

by Sumanth Kumar

PARTICLE SWARM OPTIMIZATION

CONT..
3/26/2012

14

by Sumanth Kumar

PARTICLE SWARM OPTIMIZATION

CONT..
3/26/2012

pbest is particles best position in previous generations. gbest is best position of all particles in a generation.

by Sumanth Kumar

Velocity update rule


V [i 1] V [i] c1* rand * ( pbest[i] x[i]) c2 * rand * ( gbest[i] x[i])

where c1,c2 are learning factors [0 4]

Position update rule


x[i 1] x[i] V [i 1]
15

PARTICLE SWARM OPTIMIZATION


Convergence plot for Rastrigins Funciton 82

CONT..
3/26/2012

PSO 80

by Sumanth Kumar

Value of the cost function

78

76

74

72

70

16
0 5 10 15 No of iterations 20 25

ANT COLONY OPTIMIZATION


3/26/2012

Inspired by behavior of ants in search of food. Majorly used for finding optimal paths in complex graphs. Ants lays a trail of pheromone along their path.

17

by Sumanth Kumar

ANT COLONY OPTIMIZATION

CONT..
3/26/2012

Consider a Travelling Sales Man (TSP) problem. N cities and M ants. Condition for transition from a city i to j :
Whether it is already visited or not. Distance between the cities (dij ) . Amount of artificial pheromone.

by Sumanth Kumar

Path with maximum pheromone is the shortest of all paths.

18

DIFFERENTIAL EVOLUTION (DE)


3/26/2012

Initial population candidate solutions within search space called Agents. Optimizes a problem by iteratively improving the candidate solutions. Similar to GA , DE uses recombination and mutation operators. Phases of DE :

by Sumanth Kumar

Initialization Mutation Recombination Selection

19

DIFFERENTIAL EVOLUTION (DE)

CONT..
3/26/2012

Initialization
XiG = [ X1,I,G , X2,i,G , X3,i,G, . XD,I,G ,] i= 1,2,3 N

by Sumanth Kumar

Mutation
Vi,G+1 =Xr1,G + F * (Xr2,G - Xr3,G) where F is differential weight [0 2]

Recombination V j ,i ,G 1 u j ,i ,G 1 X j ,i ,G

If randji CR or j = Irand If randji > CR or j Irand

Selection

Selected agents are passed on to next generations.


20

CONCLUSION
3/26/2012

Which Optimization algorithm is best? Which algorithm should be used for my problem?

by Sumanth Kumar

NO FREE LUNCH theorem for optimization. Every optimization algorithm is best for one or other problem. On an average all algorithms efficiency is almost equal. Simulate and find which algorithm best suites your problem.

21

REFERENCES
3/26/2012

Global Optimization Algorithms Theory and Application: Thomas Weise, Version: 2009-06-26.

Marcin Molga, Czesaw Smutnicki Test functions for optimization needs : kwietnia 2005.
Yi-Tung Kao, Erwie Zahara. "A hybrid genetic algorithm and particle swarm optimization for multimodal functions" Science Direct, Applied Soft Computing 8 (2008) 849857.

by Sumanth Kumar

Y. Shi, R. Eberhart."Empirical study of particle swarm optimization ".Proc. of Congress on Evolutionary Computation, 1999: 1945-1950.
Tutorial on Particle Swarm http://www.swarmintelligence.org/tutorials.php Optimization -

Hozefa M Botee, Eric Bonabeau: Evolving Ant Colony Optimization: Adv Computer Systems (1998) 1, 149_159. An Introduction to Differential Evolution: Kelly Fleetwood.
22

23

3/26/2012 by Sumanth Kumar

You might also like