Lecture 2: Genetic Algorithms

Introduction

Sabah Sayed
Department of Computer Science
Faculty of Computers and Artificial Intelligence
Cairo University
Egypt

2/26/2024 1
 Biological Evolution Process
“Natural Selection” + “Genetic Inheritance”

Basic Idea of Principle of Natural

Selection
“Select The Best, Discard The Rest”

2/26/2024 2
 Remember
• An individual is characterized by a set of parameters:
Genes
• The genes are joined into a string: Chromosome

• The chromosome forms the genotype

• The genotype contains all information to construct an
organism: the phenotype

• Reproduction is a “dumb” process on the chromosome

of the genotype
• Fitness is measured in the real world of the phenotype

2/26/2024 3
 What Are Genetic Algorithms (GAs)?

• Genetic Algorithms are search and

optimization techniques based on principle of
Biological Evolution Process.

• Genetic Algorithms implement optimization

strategies by simulating evolution of species
through Natural Selection.

6
 Working Mechanism Of GAs
Begin

Initialize population

T =0

Evaluate Solutions

Optimum N
Solution?
Selection
Y

T=T+1 Stop Crossover

Mutation

Replacement

7
 Simple Genetic Algorithm
Simple_Genetic_Algorithm()
{
Initialize the Population;

Calculate Fitness Function; // Evaluate Individuals …

While(Fitness Value != Optimal Value) // Continue Evolution …

{
Selection; //Natural Selection, Survival Of Fittest …

Crossover; //Reproduction, Propagate favorable characteristics …

Mutation; //Mutation …

Replacement

Calculate Fitness Function;

}
}

8
 Nature to Computer Mapping

Nature Computer

Environment Problem to be optimized.

Population Set of candidate solutions.
Individual One candidate solution to a problem.
Fitness/Survival Degree Quality of a solution.
Chromosome Encoding of a Solution to another format
(For example, bit string, floating point array).
Gene Part of the encoded solution.
Reproduction Crossover
Copying errors Mutation

9
 Components of a GA
A problem to solve, and ...
• Encoding technique (gene, chromosome)
• Initialization procedure (creation)
• Evaluation function (environment)
• Selection (reproduction, environment)
• Genetic operators (crossover, mutation)
• Parameter settings (practice and art)
 Many parameters to set
• Any GA implementation needs to decide on a
number of parameters: Population size (N),
mutation rate (m), crossover rate (c)
• Often these have to be “tuned” based on
results obtained - no general theory to deduce
good values
• Typical values might be: N = 50, m = 0.05, c =
0.9
 Example 1: A Very Simple Example
• Generate a bit string for the decimal number 31.
• Traditional Technique:
– Start by 00000 till reaching 11111
• GA Steps:
1. Build an initial population of random individuals
(chromosomes or initial solutions)
2. Evaluate fitness of individuals
3. Select some individuals to
• Apply Crossover and Mutation
4. Build next generation
5. Go to step 2.

2/26/2024 12
 Example 1: GA Step 1
• Build an initial population of random individuals

G01 G02 G03 G04

G0 00100 10100 01011 00101

• Generation 0

2/26/2024 13
 Example 1: GA Step 2
• Evaluate fitness of individuals

G01 G02 G03 G04

G0 00100 10100 01011 00101

4 20 11 5

• Fitness can be calculated as the decimal value of each bit string

2/26/2024 14
 Example 1: GA Step 3
• Select some individuals to apply crossover and mutation

G01 G02 G03 G04

G0 00100 10100 01011 00101

4 20 11 5

• Individual with good fitness should be selected for crossover and

mutation.

2/26/2024 15
 Example 1: GA Step 3
• Select some individuals to apply crossover and mutation

G01 G02 G03 G04

G0 00100 10100 01011 00101

4 20 11 5

• Individual with good fitness should be selected for crossover and

mutation.

2/26/2024 16
 Example 1: GA Step 3
• Select some individuals to apply crossover and mutation

G01 G02 G03 G04

G0 00100 10100 01011 00101

• How to apply crossover between these 2 parents?

– Single point? Multiple point? Location?
– It depends on your implementation parameters
– Crossover may happen or not according to the probability of crossover

2/26/2024 17
 Example 1: GA Step 3 – Trial 1
• Select some individuals to apply crossover and mutation

G01 G02 G03 G04

G0 00100 10100 01011 00101

G11 G12
10011 01100

• Single point crossover after 2 bits (Genes)

2/26/2024 18
 Example 1: GA Step 3 – Trial 1
• Select some individuals to apply crossover and mutation
20 11
G01 G02 G03 G04
G0 00100 10100 01011 00101

G14 G11 G12 G13

G1 00100 10011 01100 00101
4 19 12 5

Fitness?
2/26/2024 19
 Example 1: GA Step 3 – Trial 2
• Select some individuals to apply crossover and mutation

G01 G02 G03 G04

G0 00100 10100 01011 00101

G11 G12
10111 01000

• Single point crossover after 3 bits

2/26/2024 20
 Example 1: GA Step 3 – Trial 2
• Select some individuals to apply crossover and mutation
20 11
G01 G02 G03 G04
G0 00100 10100 01011 00101

G14 G11 G12 G13

00100 10111 01000 00101
4 23 8 5

Fitness?
2/26/2024 21
 Example 1: GA Step 3 – Trial 2
• Select some individuals to apply crossover and mutation
20 11
G01 G02 G03 G04
G0 00100 10100 01011 00101

G14 G11 G12 G13

00100 10111 01000 00101
4 23 8 5
• How to apply a mutation over some individual?
– For bit strings, and according to the mutation probability of your
2/26/2024
implementation, one bit might be flipped! 22
 Example 1: GA Step 3 – Trial 2
• Select some individuals to apply crossover and mutation
20 11
G01 G02 G03 G04
G0 00100 10100 01011 00101

G13 G11 G12 G14

G1 00100 10111
11111 01000 00101
4 23
31 8 5
OUR OBJECTIVE!
2/26/2024 23
 Example 1: GA Step 3 – Trial 3
• Select some individuals to apply crossover and mutation

G01 G02 G03 G04

G0 00100 10100 01011 00101

• Applying multiple point crossover!

2/26/2024 24
 G01 G02 G03 G04
G0 00100 10100 01011 00101

2/26/2024 25
 G01 G02 G03 G04
G0 00100 10100 01011 00101

2/26/2024 26
 Example 1: GA Step 3 – Trial 4
• Select ALL individuals to apply crossover and mutation
G01 G02 G03 G04
G0 00100 10100 01011 00101
4 20 11 5
G11 G12
00011 01100
3 12

2/26/2024 27
 Example 1: GA Step 3 – Trial 4
• Select ALL individuals to apply crossover and mutation
G01 G02 G03 G04
G0 00100 10100 01011 00101
4 20 11 5
G11 G13 G12 G14
00011 10101 01100 00100
3 21 12 4
G11 G13 G12 G14
00011 11101 01100 00110
3 21→29 12 4→6

2/26/2024 28
 Example 1: GA Step 4
• Build next generation .. Full Replacement or Mix?
G01 G02 G03 G04
G0 00100 10100 01011 00101
4 20 11 5
G11 G13 G12 G14
00011 10101 01100 00100
3 21 12 4
G11 G13 G12 G14
00011 11101 01100 00110
3 29 12 6

2/26/2024 29
 Example 1: GA Step 4
• Build next generation .. Mixing
G02 G03 G11 G13
G1 10100 01011 00011 11101
20 11 3 29
G02 G03 G21 G22
10100 01011 00001 11111
20 11 1 31

G02 G03 G21 G22

10100 11011 00000 11111
20 27 0 31

OUR OBJECTIVE!
2/26/2024 30
Does it really cost one iteration?
• Real world problems are not that easy.
• You could have:
– Hundreds of individuals (chromosomes)
– Huge search space
• You might perform thousands of iterations
(have thousands of generations) before
reaching your objective!

2/26/2024 31
 How do you encode a solution?

• Obviously this depends on the problem!

• GA’s often encode solutions as fixed length
“bitstrings” (e.g. 101110, 111111, 000101)
• Sometimes, your solutions (chromosomes)
will be used in their original format (floating
point numbers or alphabetical strings)
 Why does crossover work?
• The idea is that crossover preserves “good
bits” from different parents, combining
them to produce better solutions
• The fitness may be improved or get worst
after mutation
• A good scheme would therefore try to
preserve “good bits” during crossover and
mutation
 Which problems can GAs solve?
• The solution can be encoded.
• Each gene represents some aspect of the
proposed solution to the problem (the
solution can be decomposed into smaller
parts)
• The ability to “test” any solution and get a
“score” indicating how “good” that solution is
(Fitness).

2/26/2024 34