Genetic Algorithms - PowerPoint PPT Presentation

PPT – Genetic Algorithms PowerPoint presentation | free to view - id: 9f4ec-ODlkY

The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
Title:

Genetic Algorithms

Description:

Genetic Algorithms. Group Members: Leng Marcel. Tang Yong Han. Wong Leo E. This is the pirate ship. ... Algorithms. Precursor to Genetic Algorithm. Uses ... – PowerPoint PPT presentation

Number of Views:165
Avg rating:3.0/5.0
Slides: 28
Provided by: u311
Category:
Transcript and Presenter's Notes

Title: Genetic Algorithms

1
Genetic Algorithms
• Group Members
• Leng Marcel
• Tang Yong Han
• Wong Leo E

2
? This is the pirate ship.
The tiny little small boat. ? Capacity One man
and another 140kg.
3
• The Bust of Cleopatra (60kg), 40,000

4
(No Transcript)
5
(No Transcript)
6
• Discrete Knapsack Problem

Holland says The fittest strings survives
Heuristic Algorithms
According to Darwin The fittest genes survive.
Genetic Algorithms
100010111101000101001010
Optimization
7
Hill Climbing Algorithms
• Precursor to Genetic Algorithm
• Uses only one string
• Goes uphill only
• Stops when there is no higher point near itself

8
(No Transcript)
9
• Easily trapped on a local minimum
• Dependent on the initial condition

10
(No Transcript)
11
(No Transcript)
12
• So How Does GA Work ?

13

Implementation of GA
• Input of a function (Maximization)
• Sequence of a route (Traveling Salesman Problem)
• Sequence of jobs (Scheduling)
• Binary (110100101)
• Integers (2 5 4 1 3)
• n-tuples (2,4.1) (3,0.2) (1,8.9)

Representation
14
Implementation of GA
For different types of problems, we need to
choose wisely the appropriate representation.
The types of genetic operators to be used in the
GA depend on the representation, which in turn
determines the overall efficiency of the GA.
15
Evaluation Selection
• (1110100100) u1
• (1101001101) u2
• (0110001001) u3
• (0001101010) u4
• (1001010111) u5

u1 u2 u3 u4 u5
Fitness evaluation
Representation
u4 u1 u5 u4 u2
u3
u2
Selection
u1
u4
u5
16
Crossover (pc 0.40)
• (0001101010) u4
• (1110100100) u1
• (1001010111) u5
• (0001101010) u4
• (1101001101) u2

0.13 0.52 0.29 0.88 0.65
Random generation
Crossover
• (0001110111) v1
• (1110100100) v2
• (1001001010) v3
• (0001101010) v4
• (1101001101) v5

17
Mutation (pm 0.10)
• (0001110111) v1
• (1110100100) v2
• (1001001010) v3
• (0001101010) v4
• (1101001101) v5

0.03 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
0.01 _ _ _ _ _ 0.04 _ _ _ 0.09 _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _
Random generation
• (1001110111) w1
• (1110101100) w2
• (1011001110) w3
• (0001101010) w4
• (1101001101) w5

Mutation
18
Why does GA work?
• In contrast to hill-climbing, GA runs by parallel
searching.
• Schema Theorem states that short, low-order,
above-average schemata increases exponentially in
subsequent generation of a GA.

E.g. (1110100100) (11 01 )
( 100 )
• In biological terms, it means that the genes
that give rise to better survival phenotypes than
others tend to dominate the gene pool, and
finally combine to form the genotype of the best
individual.

19
The Applications of GA
Painting application
Materials handling with a robot
Robot trajectory
Robot for assembly
• Welding with
• a robot

20
The Applications of GA
• To find shapes of protein molecules

To understand protein protein docking
21
The Applications of GA
State Assignment Problem (SAP) Traveling
Salesman Problem (TSP)
22
The Applications of GA
• Game Theory (Prisoners Dilemma)
• Nonlinear dynamical Systems
• Designing neural networks
• And many others !!

23
Aspirations
24
Our trial of GA finds that
• Optimal solution to presentation
• Do not do presentation.

25
(No Transcript)
26
But due to the constraints of our grades
27
• Special thanks to
• our Mentors
• Kamalesh
• Huegesh
• Shiang Yong
• Jing Feng

Special thanks to our Mentors