Download pptx - Genetic Algorithms

Transcript
Page 1: Genetic Algorithms

GENETIC ALGORITHMSModelling nature for engineering solutions

EJJ 210Fouche, GJ13004019

Page 2: Genetic Algorithms

GENETIC ALGORITHMS Introduction Basis Advantages How the Algorithm Works Algorithm Operators Real world Applications Use in Antenna Design Conclusion

Page 3: Genetic Algorithms

INTRODUCTION – WHAT ARE GA’S ? Genetic Algorithms are a Darwinian natural-

selection based search heuristic used to ‘evolve’ solutions to abstracted problems.

First tested in 1954

Page 4: Genetic Algorithms

BASIS OF GENETIC ALGORITHMS Based on Darwin's principle of

Natural selection The idea of ‘Survival of the Fittest’ Development of advantageous attributes

through survival over generations The use of operators like Genetic crossover,

mutation and fitness evaluation to ‘evolve’ a robust solution to the problem

Information on the solution is represented as a line of ‘Genes’ in a ‘Chromosome’

Easy to code in an object orientated language

Page 5: Genetic Algorithms

ADVANTAGES OF GA’S ? Good at navigating huge search spaces, looking

for optimal combinations to form a solution to the task at hand.

Particularly useful when the solution cannot be derived mathematically

Because solutions are abstracted, genetic algorithms are applicable to many areas of engineering, intelligent and complex systems.

Solutions rated and evolved according to an objective and variable fitness criteria

Limited only by computational power. Their strongest attribute is being based on a

system that already works so perfectly – nature.

Page 6: Genetic Algorithms
Page 7: Genetic Algorithms

ALGORITHM OPERATORS - CHROMOSOME The ‘Chromosome’ is the way that

information is represented, and it contains the ‘steps’ of the solution.

In G.A. Chromosomes are simply long arrays of ‘Genes’ that encode information

Each Gene represents an encoded step, or characteristic. Often a binary representation.

Page 8: Genetic Algorithms

ALGORITHM OPERATORS & FITNESS EVALUATION

A fitness function is simply an objective measure of a Chromosomes fitness with regard to the problem at hand.

The general idea is to give a higher fitness score the closer a chromosome comes to solving the problem.

Page 9: Genetic Algorithms

APPLICATIONS OF GENETIC ALGORITHMS Stock Trading Control System debugging Evolving Medicines (IcoSystem) Airline route optimization Antenna design Efficient wind turbine design The training of Artificial

Intelligence networks Many more...

Page 10: Genetic Algorithms

USE IN ANTENNA DESIGN In the past few years, Genetic Algorithms

have been used to evolve antenna designs primarily with unusual design requirements, such as unusual radiation patterns.

Example: Antennae on the NASA ST5 satellites

Its unusual a-symmetric design would have been impossible to design using manual methods.

Page 11: Genetic Algorithms

CONCLUSION