Documente Academic
Documente Profesional
Documente Cultură
CONCEPTS OF GA
Presented By
Dr. Bhupendra Verma
Director,
Technocrats Institute of Technology( Excellence), Bhopal
ARTIFICIAL INTELLIGENCE
AI is the branch of computer science that is concerned with
the automation of Intelligent behavior.
Intelligence is not very well defined and therefore has been
less understood.
Task associated with intelligence such as learning, intuition
creativity and interference all seem to have been partially
understood.
AI in its quest to design intelligent systems has fanned out
to encompass a number of technologies in its fold.
Of these technologies NN,FL, GA are predominantly known
as Soft Computing
GENETIC ALGORITHMS
GA are
a family of computational models inspired by genetic
evolution. The basic idea is that each individual of an evolving
population encodes a candidate solution (e.g. a prediction rule) to a
given problem (e.g. classification).
GAs are adaptive, robust, efficient, and global search methods,
suitable in situations where the search space is large. They optimize
a fitness function, corresponding to the preference criterion , to
arrive at an optimal solution using certain genetic operators.
as
ADVANTAGES OF GA
Easy to understand
General purpose, robust search technique
We always get an answer and it gets better with time
Inherently parallel and easily distributed
Supports multi objective optimization
Good for noisy environment
Modular, separate from application
Simple, Powerful, Adaptive, Parallel
Guarantee near optimum solutions.
Give solutions of un-approximated form of problem.
Finer granularity of search spaces.
COMPONENTS OF A GA
Representation of individuals
String of parameters (genes) :
chromosome
eg. F(p,q,r,s,t): p q r s t
We can use bit string, arrays, trees, lists
or any other objects
Chromosome represent a candidate
solution
Bit-string representation (Ex):
100110101101100
length.
Can be easily modified for any finite
alphabets.
For example, can use 10 symbols {0,
1, . . . , 9}.
Example solution: 41662971.
Can use letters and numbers.
Example solution: THEANSWERIS42
Encoding
Hexadecimal Encoding
Permutation Encoding
( Ex Travelling Salesman Problem)
Value
Encoding
Encoding
f(x) = x2
for 2 x 1
GENETIC ALGORITHMS:
BINARY-CODED REPRESENTATIONS
ub = 2,
lb = 1,
l
= the length of the chromosome in bits
c
= the chromosome
Fitness Function
OPERATORS: SELECTION
FITNESS PROPORTIONATE SELECTION (FI/F )
NUMBER OF REPRODUCTIVE TRIALS FOR
INDIVIDUALS
Roulette-wheel selection
wheel spaced in proportion to fitness values
N (pop size) spins of the wheel
Tournament Selection
RANK SELECTION
Roulette wheel has a problem when the fitness
values of individual differs very much. If best
chromosome has fitness 90% then 90% wheel will
be occupid by the chromosome.
In Rank selection all chromosomes are assign
ranks irrespective of their fitness value.
If there are five chromosome then they will be
having proportionately 1/15 , 2/15, 3/15, 4/15 and
5/15 of the wheel
Crossover
Mutation
GENETIC ALGORITHMS:
EXAMPLE
A population size of 6,
A crossover usage rate of 1.0, and
A mutation rate of 1/7.
d(2,-2,7,c) = 4*decode(c)/127 - 2
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
Fitness
Fit: 0.109
Fit: 0.697
Fit: 1.790
Fit: 0.090
Fit: 0.238
Fit: 2.531
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
Population at t=0
Genotype Phenotype
Person 1: 1001010
0.331
Person 2: 0100101 - 0.835
Person 3: 1101010
1.339
Person 4: 0110110 - 0.300
Person 5: 1001111
0.488
Person 6: 0001101 - 1.591
Fitness
Fit: 0.109
Fit: 0.697
Fit: 1.793
Fit: 0.090
Fit: 0.238
Fit: 2.531
Is Replaced by:
Child
Child
Child
Child
Child
Child
1
2
3
4
5
6
:
:
:
:
:
:
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
Population at t=1
Genotype Phenotype Fitness
Person 1: 0001010
- 1.685 Fit: 2.839
Person 2: 1101101
1.433 Fit: 2.054
Person 3: 1011100
0.898 Fit: 0.806
Person 4: 0001011
- 1.654 Fit: 2.736
Person 5: 1101010
1.339 Fit: 1.793
Person 6: 1010101
0.677 Fit: 0.458
GENETIC ALGORITHMS:
AN EXAMPLE RUN (BY HAND)
CONVERGENCE
Population
Schema
HYPER-PLANE MODEL
Search space
Individuals
Schemas
0**
SAMPLING HYPER-PLANES
Each vertex
Member of 3L hyper-planes
Samples hyper-planes
SCHEMA THEOREM
FORMAL STATEMENT
Selection probability
Crossover probability
m( H , t ) f ( H , t )
E (m( H , t 1))
f
(H )
c
L 1
P(hcrossover ) p
Mutation probability
P (hmutation ) ( H ) pm
Expected number of members of a
schema m( H , t ) f ( H , t )
(H )
E (m( H , t 1)
(1 pc
L 1
pm( H ))
AREA OF APPLICATION
GAs can be used when:
Non-analytical problems.
Non-linear models.
Uncertainty.
Large state spaces.
NON-ANALYTICAL PROBLEMS
NON-LINEAR MODELS
UNCERTAINTY
Changing parameters.
State space
understood.
may
not
be
completely
GA APPLICATION EXAMPLES
Function optimizers
difficult, discontinuous, multi-modal, noisy functions
Combinatorial optimization
layout of VLSI circuits, factory scheduling, traveling salesman
problem
Design and Control
bridge structures, neural networks, communication networks
design; control of chemical plants, pipelines
Machine learning
classification rules, economic modeling, scheduling strategies
Portfolio design, optimized trading models, direct marketing
models, sequencing of TV advertisements, adaptive agents, data
mining, etc
Constrained mathematical
optimization problems especially
when there are few solutions.
Constraints are difficult to
incorporate into a GA.
Guided domain search is possible
and efficient.