You are on page 1of 46

Applications of

Artificial Intelligence
Techniques in Engineering

Dr. Anil Swarnkar


Assistant Professor
Department of Electrical Engineering
Malaviya National Institute of Technology Jaipur
Email: mnit.anil@gmail.com
Artificial Intelligence
Techniques
Guide an underlying search to escape from being trapped
in a local optima and to explore better areas of the
solution space

Single solution approaches


Concerning only one solution at a particular time during
the search. Search is very much restricted to local regions,
so called local
•Simulated Annealing,
•Tabu Search,
•variable neighbourhood search, etc.
Population based approaches
concern a population of solutions at a time
•Genetic algorithm
•Particle Swarm Optimisation
•Ant Colony Optimisation, etc.
Maxima and Minima
Rastrigin's Function
R(x) = 20 + x12 + x22 -10(cos 2πx1 + cos 2 πx2
Ackley’s function
How AI works
Initialization
Generation/Iteration 60
Generation/Iteration 80
Generation/Iteration 95
Generation/Iteration 100
Genetic
Algorithms
Genetic Algorithm: Basic Idea

Charles Darwin
Origin of the Species
Survival of the Fittest
1859
History of Genetic Algorithm

 Genetic algorithms belong to the


larger class of evolutionary
algorithms (EA)
 Idea of evolutionary computing was
introduced in the 1960s by I.
Rechenberg in his work "Evolution
strategies" .
 His idea was then developed by
other researchers. Genetic
Algorithms (GAs) were invented by
John Holland and developed by him
and his students and colleagues.
This lead to Holland's book
"Adaption in Natural and Artificial
Systems" published in 1975.
History of Genetic Algorithm
 In 1992 John Koza has used
genetic algorithm to evolve
programs to perform certain
tasks. He called his method
"genetic programming" (GP).
 The development of GA is
also credited to the work of
D. E. Goldberg published in
“Genetic algorithms in
search, optimization and
machine learning”.
Genetic Algorithm:
Methodology
 The Genetic Algorithms are
derivative free stochastic
optimization methods inspired by
Darwin's theory of the concept of
natural selection and evolution.
 A genetic algorithm (GA) is a search
heuristic which generate solutions to
optimization problems using
techniques inspired by natural
evolution, such as selection,
crossover and mutation.
Terminology
Chromosome:
All living organisms consist of
cells. In each cell there is the
same set of chromosomes.
Chromosomes are strings of
DNA (Deoxyribonucleic Acid)
and serves as a model for the
whole organism.
A chromosome consists of
genes, blocks of DNA. Each
gene encodes a trait, for
example color of eyes.
Terminology
Chromosome:
In GA, the chromosome may be defined
as the genetic representation of the
problem solution. A chromosome is a set
of parameters which encode a candidate
solution (also called individuals) of the
problem to be solved by GA. The
chromosome is often represented as a
simple string composed of real or binary
numbers.
Encoding
Phenotype space Genotype space
Encoding
(representation) 10010001
10010010
010001001
011101001
Decoding
(inverse representation)
Terminology
Fitness Function:
The fitness function is defined based on
the optimization problem, which
measures the quality of the represented
solution.
The fitness is evaluated using the set of
parameters defined in the chromosome
string of an individual using the objective
function of the problem
Terminology
Initialization:
Initially many individuals are
randomly generated to form the
initial population covering the entire
range of possible solutions (the
search space).
The population size depends on the
nature of the problem and the
search space.
Terminology
Selection:
During the evolutionary process of GA
the parents are selected from the
current generation to produce the
descendants for next generation.
There are many methods available in
the literature for the selection of
parents. Some of the commonly used
methods are roulette wheel,
tournament, etc.
Roulette Wheel Selection
Parents are selected according to their
fitness. The better the chromosomes are,
the more chances to be selected they have.
Imagine a roulette wheel where are placed
all chromosomes in the population, every
has its place big accordingly to its fitness
function, like on the following picture.
Tournament selection
Tournament selection involves running
several "tournaments" among a few
individuals chosen at random from the
population.
The winner of each tournament (the
one with the best fitness) is selected
for crossover.
Selection pressure is easily adjusted
by changing the tournament size.
If the tournament size is larger, weak
individuals have a smaller chance to
be selected.
Reproduction

Elitism

Crossover/recombination

Mutation
One Point Crossover
N-Point Crossover
Uniform Crossover
Mutation

 Changing one/more
bit(s) at randomly
selected position(s)
of a parent.
 Small probability
(usually 2-3 % of
popsize)
Flow Chart of Genetic
Algorithm
Particle Swarm
Optimisation
Particle Swarm Optimisation
 Particle Swarm Optimization (PSO), inspired by
social behavior of bird flocking or fish schooling, is a
population based stochastic optimization technique
developed by Eberhart and Kennedy in 1995
Position Update
A swarm consists of number of particles “or possible
solutions” that fly through the feasible solution space to
explore optimal solutions. Each particle updates its position
based on its own best exploration; best swarm overall
experience, and its previous velocity vector according to the
following model:
pbest j  s kj gbest j  s kj (1)
v kj 1  Wv kj  C1  rand1 ()   C2  rand2 () 
t t
s kj 1  s kj  v kj 1  t (2)

First term represents the „inertia‟ of the particle, second term


represents the „personal influence‟ and the last term represents the
„social influence‟ of the particle. Equation (2) finds the new position
of the particle according to the new velocity calculated by (1).
Position Update

 W  Wmax 
W (k )   min  (k  1)  Wmax
 itermax  1 
Concept of multi-agent search by
PSO in a multi-dimensional space
Ant Colony
Optimisation
Ant Colony Optimization

The ACO was initially proposed by


Marco Dorigo in 1992. The search
technique is inspired by the behavior
of ants in finding paths from the nest
to food and back to the nest without
any visual clues.
Ant Intelligence
Mathematical Formulation
The ant communication is accomplished primarily through the
chemicals called pheromones. While moving, the ants deposit
the pheromone trail on the ground. Other ants perceive the
presence of the pheromone and tend to follow paths where the
pheromone concentration is higher. The probability with which
the kth ant will move from node i to node j can be determined by
the random-proportional state transition rule as given below:

where τi,j is the amount of the pheromone on the edge i-j, α is the
parameter to control the influence of τi,j, ηi,j is the desirability of
ant to move from node i to node j based on a priori knowledge,
typically 1/di,j, (where dij is the distance between node i and
node j) and β is the parameter to control the influence of ηi,j.
Pheromone Update

where τi,j is the amount of pheromone on a given edge i-j, ρ is the


pheromone evaporation rate and

is the amount of pheromone deposited by the kth ant,


typically given by
Comparison
S. Feature GA PSO ACO
No.
1. Evolutionary Crossover and Mutation Velocity update using pbest Pheromone and Desirability
operator and gbest
2. Binary/Real Can be designed in both Can be designed in both Can be designed in both binary and real
binary and real form binary and real form form
3. Control Crossover rate (CR), Acceleration coefficients (C1, Pheromone control parameter (α),
parameters Mutation rate (MR) C2), Inertia weight (W) Desirability control parameter (β), Hunting
group size (ζ), Evaporation (ρ) and Global
update constant multiplier (σ)
4. Escaping from Using mutation With the help of pbest & gbest Evaporation mechanism
local optimum
5. Complexity Complex computation Easy computation Complex computation to manage
pheromone concentration and desirability
for each edge.
6. Memory Less Less Large Memory Requirement to store
requirement Pheromones concentration and desirability
7. Survival of the Applicable Not applicable Not applicable
fittest
8. Convergence Fast Slow Fast
9. Population size More Less Less
required
10 Continuous Applicable to both Mainly devised for continues Applicable to both continues and discrete
/Discrete continues and discrete variables but can be applied to variable problems
variable problems continuous variable problems
with approximation
Other AI Techniques
 Cockoo Search
 Honey Bee Mating Optimisation
 Biogeography-based Optimisation
 Artificial Neural Network
 Immune Algorithm
 etc…
Applications
 Electrical Engineering
◦ Distribution Network Reconfiguration
◦ Optimal Capacitor Placement
◦ Economic Load Dispatch
◦ Optimal Penetration of Distributed
Generation
◦ Optimisation of Machine Parameters
◦ Optimal Placement of FACTs in
Transmission System
Applications
 Electronics & Communication
Engg.
 Mechanical Engineering
 Civil Engineering
 Chemical Engineering
 and what not.....
Conclusion
 All these techniques are capable of exploring
the global or near global solution of complex
combinatorial problems.
 However, they suffer from the drawback of
heavy computational burden
 Their convergence and quality of solutions
largely depend upon the appropriate choice of
their control parameters and quality of
population space or search space.
Thank
You

You might also like