Ga PDF

R.K.
Bhattacharjya/CE/IITG
Introduction To Genetic Algorithms
Dr. Rajib Kumar Bhattacharjya

Department of Civil Engineering
IIT Guwahati
Email: rkbc@iitg.ernet.in
19 October 2012
References
R.K. Bhattacharjya/CE/IITG
D. E. Goldberg, Genetic Algorithm In

Search, Optimization And Machine Learning, New
York: Addison Wesley (1989)
John H. Holland Genetic Algorithms, Scientific

American Journal, July 1992.
Kalyanmoy Deb, An Introduction To Genetic

Algorithms, Sadhana, Vol. 24 Parts 4 And 5.
19 October 2012
Introduction to optimization
Global optima
Local optima
Local optima
Local optima
Local optima
19 October 2012
Introduction to optimization
Multiple optimal solutions
19 October 2012
Genetic Algorithms
5
Genetic Algorithms are the heuristic search

and optimization techniques that mimic the
process of natural evolution.
19 October 2012
Principle Of Natural Selection

6
Select The
Best,
Discard The
Rest
19 October 2012
An Example.
Giraffes have long necks

Giraffes with slightly longer necks could feed on

leaves of higher branches when all lower ones had
been eaten off.
They had a better chance of survival.
Favorable characteristic
propagated through generations
of giraffes.
Now, evolved species has long necks.
19 October 2012
An Example.
8
This longer necks may have due to the effect of mutation

initially. However as it was favorable, this was propagated over
the generations.
19 October 2012
Evolution of species
Initial Population of
animals
Millions Of
Years
Struggle For Existence

Survival Of the Fittest
Surviving Individuals
Reproduce, Propagate Favorable
Characteristics
Evolved Species
19 October 2012
10
Thus genetic algorithms implement

the optimization strategies by
simulating evolution of species
through natural selection
19 October 2012
Simple Genetic Algorithms

11
Start
Initialize
population
Evaluate Solutions
T =0
Optimum
Solution?
Selection
Y
T=T+1
Stop
Crossover
Mutation
19 October 2012
Simple Genetic Algorithm

12
function sga()
{
Initialize population;
Calculate fitness function;
While(fitness value != termination criteria)
{
Selection;
Crossover;
Mutation;
Calculate fitness function;
}
}
19 October 2012
Nature to Computer Mapping

13
Nature
Computer
Population
Individual
Fitness
Chromosome
Gene
Reproduction
Set of solutions
Solution to a problem
Quality of a solution
Encoding for a solution
Part of the encoding solution
Crossover
19 October 2012
GA Operators and Parameters

14
Selection
Crossover
Mutation
Elitism
19 October 2012
Encoding
15
The process of representing a solution in the

form of a string that conveys the necessary
information.
Just as in a chromosome, each gene controls a
particular characteristic of the
individual, similarly, each bit in the string
represents a characteristic of the solution.
19 October 2012
Encoding Methods
16
Most common method of encoding is binary coded.

Chromosomes are strings of 1 and 0 and each position
in the chromosome represents a particular characteristic
of the problem
Decoded value
52
26
Mapping between decimal

and binary value
19 October 2012
Encoding Methods
17
Defining a string
[0100001010]
d
h
(d,h) = (8,10) cm
Chromosome = [0100001010]
19 October 2012
Fitness function
18
A fitness function value quantifies the optimality

of a solution. The value is used to rank a
particular solution against all the other solutions.
A fitness value is assigned to each solution
depending on how close it is actually to the
optimal solution of the problem.
19 October 2012
Assigning a fitness function

19
Considering c = 0.0654
23
19 October 2012
Selection
20
The process that determines which solutions are

to be preserved and allowed to reproduce and
which ones deserve to die out.
The primary objective of the selection operator is
to emphasize the good solutions and eliminate
the bad solutions in a population while keeping
the population size constant.
Selects the best, discards the rest
19 October 2012
Selection
21
Identify the good solutions in a population.
Make multiple copies of the good solutions.
Eliminate bad solutions from the population so

that multiple copies of good solutions can be
placed in the population.
19 October 2012
Selection
22
There are different techniques to implement

selection in Genetic Algorithms.
They are:
Tournament
selection
Roulette wheel selection
Rank selection
Steady state selection, etc
19 October 2012
Tournament selection
23
In tournament selection several tournaments

are played among a few individuals. The
individuals are chosen at random from the
population.
The winner of each tournament is selected
for next generation.
Selection pressure can be adjusted easily by
changing the tournament size.
Weak individuals have a smaller chance to be
selected if tournament size is large.
19 October 2012
Tournament selection
24
32
22
22
25
25
13 +30
32
40
22
32
40
32
25
22
22
25
7 +45
7 +45
25
25
22
13 +30
13 +30
19 October 2012
13 +30
Roulette wheel selection

25
Chromosome #
Parents are
selected
according to their
fitness values.
The better
chromosomes
have more
chances to be
selected.
Chromosome 1
Chromosome 2
Chromosome 3
Chromosome 4
Chromosome 5
Chromosome 6
Fitness
% of Roulette wheel
127.50
50.00
25.00
20.00
15.00
12.50
5%
51
20
10
8
6
5
Roulette Wheel
6%
8%
Chromosome 1
Chromosome 2
10%
51%
Chromosome 3
Chromosome 4
Chromosome 5
Chromosome 6
20%
19 October 2012
Rank selection
26
Chromosome # Fitness
The Roulette wheel

selection will have
problems when the
fitness values differ very
much.
In this case, the
chromosomes with low
fitness values will have a
very few chance to be
selected.
This problem can be
avoided using ranking
selection.
Chromosome 1 127.50
Chromosome 2 50.00
Chromosome 3 25.00
Chromosome 4 20.00
Chromosome 5 15.00
Chromosome 6 12.50
5%
% of
Rank
% of
Roulette
Roulette
wheel
wheel
51
6
28
20
5
24
10
4
19
8
3
14
6
2
10
5
1
5
Rank Selection
10%
28%
Chromosome 1
14%
Chromosome 2
Chromosome 3
Chromosome 4
Chromosome 5
19%
Chromosome 6
24%
19 October 2012
Steady state selection

27
In this method, a few good chromosomes are

used for creating new offspring in every
iteration.
Then some bad chromosomes are removed and
the new offspring is placed in their places.
The rest of population migrates to the next
generation without going through selection
process.
19 October 2012
Binary Crossover
28

The crossover operator is used to create new solutions from the

existing solutions available in the mating pool after applying
selection operator.
This operator exchanges the gene information between the
solutions in the mating pool.
The most popular crossover selects any two solutions strings
randomly from the mating pool and some portion of the strings is
exchanged between the strings.
The selection point is selected randomly.
A probability of crossover is also introduced in order to give
freedom to an individual solution string to determine whether the
solution would go for crossover or not.
19 October 2012
Binary Crossover
29
Source: Deb 1999

19 October 2012
Binary Mutation
30

Mutation is the occasional introduction of new features in to the solution

strings of the population pool to maintain diversity in the population.
Though crossover has the main responsibility to search for the optimal
solution, mutation is also used for this purpose.
Mutation operator changes a 1 to 0 or vise versa, with a mutation
probability of .
The mutation probability is generally kept low for steady convergence.
A high value of mutation probability would search here and there like a
random search technique.
19 October 2012
Source: Deb 1999
Elitism
31
Crossover and mutation may destroy the best

solution of the population pool
Elitism is the preservation of few best solution
of the population pool
Elitism is defined in percentage or in number
19 October 2012
An example problem
32
Consider 5 bit string to represent the solution, then

00000 = 0 and 11111 =
Let assume a population size of 4
19 October 2012
An example problem
33
Source: Deb 1999

19 October 2012
Real coded Genetic Algorithms

34
Disadvantage of binary coded GA

more
computation
lower accuracy
longer computing time
solution space discontinuity
hamming cliff
19 October 2012

35
The standard genetic algorithms has the following steps

1.
2.
3.
4.
5.
6.
Choose initial population

Assign a fitness function
Perform elitism
Perform selection
Perform crossover
Perform mutation
In case of standard Genetic Algorithms, steps 5 and

6 require bitwise manipulation.
19 October 2012

36
Simple crossover: similar to binary crossover
P1=[8 6 3 7 6]
P2=[2 9 4 8 9]
C1=[8 6 4 8 9]
C2=[2 9 3 7 6]
19 October 2012

37
Linear Crossover
Parents: (x1,,xn ) and (y1,,yn )

Select a single gene (k) at random
Three children are created as,
( x1 , ..., xk , 0.5 yk + 0.5 xk , ..., xn )

( x1 , ..., xk , 1.5 yk 0.5 xk , ..., xn )
( x1 , ..., xk , - 0.5 yk + 1.5 xk , ..., xn )
From the three children, best two are selected for the
next generation
19 October 2012

38
Single arithmetic crossover

Select a single gene (k) at random
child1 is created as,
( x1 , ..., xk , yk + (1 ) xk , ..., xn )
reverse for other child. e.g. with = 0.5
0.5 0.7 0.7 0.5 0.2 0.8 0.3 0.9 0.4
0.5 0.7 0.7 0.5 0.2 0.5 0.3 0.9 0.4
0.1 0.3 0.1 0.3 0.7 0.2 0.5 0.1 0.2
0.1 0.3 0.1 0.3 0.7 0.5 0.5 0.1 0.2

19 October 2012

39
Simple arithmetic crossover

Pick random gene (k) after this point mix values
child1 is created as:
( x , ..., x , y
+ (1 ) x
, ..., y + (1 ) x )
1
k
k +1
k +1
n
n
0.5 0.7 0.7 0.5 0.2 0.8 0.3 0.9 0.4
0.5 0.7 0.7 0.5 0.2 0.5 0.4 0.5 0.3
0.1 0.3 0.1 0.3 0.7 0.2 0.5 0.1 0.2
0.1 0.3 0.1 0.3 0.7 0.5 0.4 0.5 0.3

19 October 2012

40
Whole arithmetic crossover
Most commonly used

child1 is:
x + (1 ) y
0.5 0.7 0.7 0.5 0.2 0.8 0.3 0.9 0.4
0.3 0.5 0.4 0.4 0.4 0.5 0.4 0.5 0.3
0.1 0.3 0.1 0.3 0.6 0.2 0.5 0.1 0.2
0.3 0.5 0.4 0.4 0.4 0.5 0.4 0.5 0.3

19 October 2012
Simulated binary crossover

41
Developed by Deb and Agrawal, 1995)
Where,
a random number
is a parameter that controls the crossover process. A high value
of the parameter will create near-parent solution
19 October 2012
Random mutation
42
Where
is a random number between [0,1]
Where,
is the user defined maximum
perturbation
19 October 2012
Normally distributed mutation

43
A simple and popular method
Where
is the Gaussian probability

distribution with zero mean
19 October 2012
Polynomial mutation
44
19 October 2012
Multi-modal optimization
45
19 October 2012
After Generation 200

46
19 October 2012
Multi-modal optimization
47
Sharing function
Niche count
Modified fitness
19 October 2012
Hand calculation
48
Maximize
Sol
String
Decoded
value
110100
52
1.651
0.890
101100
44
1.397
0.942
011101
29
0.921
0.246
001011
11
0.349
0.890
110000
48
1.524
0.997
101110
46
1.460
0.992
19 October 2012
Distance table
49
dij
0.254
0.73
1.302
0.127
0.191
0.254
0.476
1.048
0.127
0.063
0.73
0.476
0.572
0.603
0.539
1.302
1.048
0.572
1.175
1.111
0.127
0.127
0.603
1.175
0.064
0.191
0.063
0.539
1.111
0.064
19 October 2012
Sharing function values

50
sh(dij)
nc
0.492
0.746
0.618
2.856
0.492
0.048
0.746
0.874
3.16
0.048
1.048
0.746
0.746
0.872
3.364
0.618
0.874
0.872
3.364
19 October 2012
Sharing fitness value

51
Sol
String
Decoded
value
110100
52
1.651
0.890
101100
44
1.397
0.942
011101
29
0.921
0.246
001011
11
0.349
0.890
110000
48
1.524
0.997
101110
46
1.460
0.992
nc
f
2.856 0.312
3.160 0.300
1.048 0.235
1.000 0.890
3.364 0.296
3.364 0.295
19 October 2012
52
Solutions obtained using modified

fitness value
19 October 2012
Evolutionary Strategies
53
ES use real parameter value
ES does not use crossover operator
It is just like a real coded genetic algorithms with

selection and mutation operators only
19 October 2012
Two members ES: (1+1) ES

54
In each iteration one parent is used to create

one offspring by using Gaussian mutation
operator
19 October 2012

55

Step1: Choose a initial solution and a mutation

strength
Step2: Create a mutate solution
Step 3: If
, replace
with
Step4: If termination criteria is satisfied, stop, else
go to step 2
19 October 2012

56
Strength of the algorithm is the proper value of
Rechenberg postulate
The
ratio of successful mutations to all the mutations

should be 1/5. If this ratio is greater than 1/5, increase
mutation strength. If it is less than 1/5, decrease the
mutation strength.
19 October 2012

57
A mutation is defined as successful if the mutated

offspring is better than the parent solution.
If
is the ratio of successful mutation over n trial,
Schwefel (1981) suggested a factor
in
the following update rule
19 October 2012
Matlab code
58
19 October 2012
59
19 October 2012
Some results of 1+1 ES

60
5
0
10
20 0
4.5
4
50
10 0
Optimal Solution is
X*= [3.00
1.99]
50
10 0
3.5
20
5
0
10
1.5
10
20
2
10
20
50
20
10 0
10 0
20 0
0.5
1.5
Objective function value f

= 0.0031007
50
20
0
0.5
50
20
10
2.5
20 0
2.5
X
3.5
4.5
20 0
19 October 2012
Multimember ES
61
ES
Step1: Choose an initial population of
and mutation strength
Step2: Create mutated solution
solutions
Step3: Combine
and , and choose the best
solutions
Step4: Terminate? Else go to step 2
19 October 2012
Multimember ES
62
ES
Through selection
Through mutation
19 October 2012
Multimember ES
63
Offspring
ES
Through selection
Through mutation
19 October 2012
Multi-objective optimization
Comfort
64
Price
19 October 2012
65
Two objectives are

Minimize weight
Minimize deflection
19 October 2012
66

More than one objectives

Objectives are conflicting in nature
Dealing with two search space
Decision variable space
Objective space

Unique mapping between the objectives and often the

mapping is non-linear
Properties of the two search space are not similar
Proximity of two solutions in one search space does not
mean a proximity in other search space
19 October 2012
67
19 October 2012
NSGA II
68
Non-dominated Sorting Genetic Algorithms

NSGA
II is an elitist non-dominated sorting Genetic

Algorithm to solve multi-objective optimization problem
developed by Prof. K. Deb and his student at IIT
Kanpur.
It has been reported that NSGA II can converge to the
global Pareto-optimal front and can maintain the
diversity of population on the Pareto-optimal front
19 October 2012
Non-dominated sorting
69
1
Feasible Region
5
4
6
2
3
Objective 1 (Minimize)
1
2
Infeasible Region
19 October 2012
Initialize population of size N

Calculate all the objective functions
Rank the population according to nondominating criteria
70
Selection
No
Crossover
Termination
Criteria?
Yes
Mutation
Calculate objective function of the
new population
Pareto-optimal solution
Combine old and new population
Non-dominating ranking on the

combined population
Calculate crowding distance of all
the solutions
Get the N member from the
combined population on the basis
of rank and crowding distance
Replace parent population by the better
members of the combined population
19 October 2012
Calculation crowding distance

71
19 October 2012
Replacement scheme of NSGA II

72
Non-dominating sorting
Pt
Pt+1
F1
F2
F3
Crowding distance
sorting
Qt
Rejected
19 October 2012
73
THANKS
19 October 2012

Ga PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ga PDF

Uploaded by

Copyright:

Available Formats

R.K.

Introduction To Genetic Algorithms

Dr. Rajib Kumar Bhattacharjya

D. E. Goldberg, Genetic Algorithm In

John H. Holland Genetic Algorithms, Scientific

Kalyanmoy Deb, An Introduction To Genetic

Multiple optimal solutions

Genetic Algorithms are the heuristic search

Principle Of Natural Selection

Giraffes have long necks

Giraffes with slightly longer necks could feed on

This longer necks may have due to the effect of mutation

Struggle For Existence

Thus genetic algorithms implement

Simple Genetic Algorithms

Simple Genetic Algorithm

Nature to Computer Mapping

GA Operators and Parameters

The process of representing a solution in the

Most common method of encoding is binary coded.

Mapping between decimal

A fitness function value quantifies the optimality

Assigning a fitness function

The process that determines which solutions are

Identify the good solutions in a population.

Make multiple copies of the good solutions.

Eliminate bad solutions from the population so

There are different techniques to implement

In tournament selection several tournaments

Roulette wheel selection

The Roulette wheel

Steady state selection

In this method, a few good chromosomes are

The crossover operator is used to create new solutions from the

Source: Deb 1999

Mutation is the occasional introduction of new features in to the solution

Source: Deb 1999

Crossover and mutation may destroy the best

Consider 5 bit string to represent the solution, then

Let assume a population size of 4

Source: Deb 1999

Real coded Genetic Algorithms

Disadvantage of binary coded GA

Real coded Genetic Algorithms

The standard genetic algorithms has the following steps

Choose initial population

In case of standard Genetic Algorithms, steps 5 and

Real coded Genetic Algorithms

Simple crossover: similar to binary crossover

Real coded Genetic Algorithms

Parents: (x1,,xn ) and (y1,,yn )

( x1 , ..., xk , 0.5 yk + 0.5 xk , ..., xn )

( x1 , ..., xk , - 0.5 yk + 1.5 xk , ..., xn )

Real coded Genetic Algorithms

Single arithmetic crossover

Parents: (x1,,xn ) and (y1,,yn )

reverse for other child. e.g. with = 0.5

0.5 0.7 0.7 0.5 0.2 0.8 0.3 0.9 0.4

0.5 0.7 0.7 0.5 0.2 0.5 0.3 0.9 0.4

0.1 0.3 0.1 0.3 0.7 0.2 0.5 0.1 0.2

0.1 0.3 0.1 0.3 0.7 0.5 0.5 0.1 0.2

Real coded Genetic Algorithms

Simple arithmetic crossover