Evolutionary Programming

Applications to Electrical
Systems
What is an optimization technique?
● Optimization is the process of adjusting the inputs to find the maximum or minimum value of a
function or process.
What is a genetic algorithm and why is it useful?
● Genetic Algorithm is an optimization method based on mechanics of natural selection and natural
genetics. Its fundamental principle is that the fittest member of population has highest probability of
survival (Based on Darwin’s theory).
● GA works with only objective function information in search for an optimal parameter set.
● GA comparatively doesn’t involve complications usually associated with calculus based schemes.
Characteristics of Genetic Algorithm
● The GA works on the coding of the parameters set rather than the actual parameters.
● The GA searches for optimal points using a population of possible solution points not a single point.
This is an important characteristic which makes GA more powerful and also results into implicit
parallelism.
● GA uses only objective function information and no other derivatives needed.
● GA uses probabilistic transition rules not the deterministic rules.
Components and Mechanism of Genetic Algorithm
● Generate randomly the initial population
● Select the chromosomes with best fitness values
● Recombine selected chromosomes using crossover and mutation operators
● Insert Offspring into the population.
● If a stop criterion is satisfied ,return the chromosomes with the best fitness. Otherwise
go to Step 2
Step 1- Representation
● The representation of chromosomes in GA has very deep impact on performance of GA based
function . There are different methods to represent like binary encoding ,value encoding
,permutation encoding proposed by Holland. A bit string encoding contains 0 or 1
● Each Chromosome of population consists of same length of binary string.

Chromosome Value Binary Encoding

X 8 01000

Y 6 00110

Here the X and Y value are 8 and 6 respectively with string length of 5
Step 2- Initial Population
● GA does not work with single string but with a population strings which evolves iteratively by
generating new individuals taking the place of their parents.To begin,the initial population is
generated at random or through use of specified information.
● Method 1- Random initialization - Populate the initial population with completely random solutions.
● Method 2 - Heuristic initialization- Populate the initial population using a known heuristic for the
problem.But its recommended the entire population should not be initialized using heuristic as it can
result in population having similar solutions and very little diversity.
● Random solutions can drive the population to optimality . It is the diversity that can lead to
optimality.
● Steady State Population model- In steady state GA,we generate one or two off springs in each
generation and they replace one or two individuals from the population.
● Generational- we generate ‘n’ offsprings where n is population size replaced by new one at end of
iteration.
Step 3 -Selection
● Selection is simply an operation whereby an old string is copied into a mating pool according to its
fitness.The selection operator selects chromosomes from the current generation to be parents for
next generations. The problem is how to select the chromosome individuals . According to Darwin’s
theoryof evolution,Best ones survive to create new offspring.
● Uniform Selection- In this selection method ,parents are selected at random from a uniform
distribution using the expectations and number of parents. This results in an undirected search.
This is not a useful search strategy , but can be used to test genetic algorithm
● Roulette wheel- This is a stochastic method where individual having higher fitness value have more
chances to select. 1) do summation of fitness of all chromosomes. 2) Generate random number R
between 0 and S.3) Go through population and sum the fitness from 0. When S greater than R ,
stop and return chromosome. 4)Repeat process until number of chromosomes= size of population
Step 3 -selection(contd)
Tournament Selection
Tournament selection provides selection pressure by holding a tournament among s competitors where s
is tournament. Winner of tournament is individual with highest fitness of s competitors. Then winner
inserted into mating pool. The mating pool comprise of tournament winner,has a higher fitness value than
average population fitness.
Step 4 - Crossover
Crossover is a randomized yet structured recombination approach . Simple crossover may proceed in
two steps. Firstly the newly reproduced strings in mating pool are mated at random. Secondly the
crossover of each pair of strings undergoes crossovers.
1. Single point crossover - crossover position is randomly selected between one and L-1 where L is
length of chromosome and two parents are crossed at that point.

2. Double Crossover
Step 5- Mutation
● Mutations are global searches . A probability of mutation is again pre determined before the algorith
is started which is applied to each individual bit of each spring chromosome to see if it is inverted .
As new individuals are generated each character generated with given probability Pm
● In simple GA mutation means a 1 to a 0 and vice versa with mutation probability
● Before Mutation-010100010010
● After Mutation- 011100000010.
● Mutation probability Pm controls rate at which new gene values are introduced into population. If
too small then many gene values that would have been useful are never tried out, Too high and off
spring lose resemblance.
● Mutation is appropriately the secondary mechanism of genetic algorithm adaptation.
Evolutionary Algorithms - Applications

● Power Flow Analysis