Robotics 16

download Robotics 16

of 13

Transcript of Robotics 16

  • 8/8/2019 Robotics 16

    1/13

    EXPERT SYSTEMS AND SOLUTIONS

    Email: [email protected]

    [email protected]

    Cell: 9952749533www.researchprojects.info

    PAIYANOOR, OMR, CHENNAI

    Call For Research Projects Final

    year students of B.E in EEE, ECE, EI,

    M.E (Power Systems), M.E (Applied

    Electronics), M.E (Power Electronics)

    Ph.D Electrical and Electronics.

    Students can assemble their hardware in our

    Research labs. Experts will be guiding theprojects.

  • 8/8/2019 Robotics 16

    2/13

    SymbolicArtificial

    Intelligence

    Lecture 10:

    Geneticalgorithms

    (continued)

    Solution

    to

    previousexercise:

  • 8/8/2019 Robotics 16

    3/13

    Selection methods

    There are many ways

    to select chromosomes

    to survive to the next

    generation

    Roulette wheelselection: the better the

    chromosome, the more

    chance for selection it

    possesses; imagine a

    roulette wheel whereevery chromosome is

    represented in

    proportion to its fitness

    function

    Then a roulette ball is thrown

    and selects chromosomes

    chromosomes with bigger fitness

    will be selected more times (if

    duplicate chromosomes are

    allowed).

  • 8/8/2019 Robotics 16

    4/13

    Other selection methods

    Rank selection: Roulette wheel selection does not work ifone chromosome has a fitness far in excess of the others;

    Rank selection is better in these cases; rank selection first

    ranks the population and then every chromosome receives

    a fitness from this ranking (the method used in ourprevious examples);

    Steady state selection: in every generation the best

    chromosomes are selected for mutation and crossover, the

    worst removed, and the remainder survive to the newgeneration;

    Elitism: the best chromosomes always survive and the rest

    are discarded - prevents losing good chromosomes.

  • 8/8/2019 Robotics 16

    5/13

    Encoding methods

    Chromosome A: 10110101010001010101

    Chromosome B: 01010001010100101000

    Encoding techniques (ways of representing the chromosome)

    depend on the problem

    binary encoding: the most commonly used - every

    chromosome is a string of bits of 1s and 0s:

    Binary encoding is efficient but not always natural;

    sometimes corrections must be made after crossover andmutation to ensure that the genotype means something at the

    phenotype level

  • 8/8/2019 Robotics 16

    6/13

    Other encoding techniques

    permutation encoding: every chromosome is a string of

    numbers which represents a number in a sequence (the method

    used in our examples in previous lecture):chromosome A: 1 5 3 2 6 5

    chromosome B: 5 3 6 2 4 7

    Again, corrections may be required after mutation and

    crossover

    value encoding: chromosomes can consist of different

    types of value (e.g. real numbers, characters):

    chromosome A: 1.232 3.45 2.65 0.454

    chromosome B: ABDDDHSGHGSHGSGSHGSWE

    chromosome C: (back) (right) (left) (forward)

    Value encoding is useful for certain specialist problems (e.g.

    evolving weights for neural networks), but requires special

    mutation and crossover mechanisms

  • 8/8/2019 Robotics 16

    7/13

    Crossover mechanisms

  • 8/8/2019 Robotics 16

    8/13

    Mutation methods

    bit inversion, e.g. 1000000001 =1010000000 where the third and

    10th bits have been (randomly) mutated;

    order changing, e.g. (5 6 3 4 7 3) = (5 3 4 6 7 3), where the

    second, third and fourth values have been randomly scrambled;

    value changing, e.g. (3.4 4.2 4.6 6.4 3.2) = (3.4 4.2 4.5 6.4 3.2)

    where one value has been changed within a specific range

  • 8/8/2019 Robotics 16

    9/13

    Further reading and references

    Genetic algorithm toolbox for Matlab.

    http://www.shef.ac.uk/~gaipp/ga-toolbox/

    Tutorials with code:

    http://evonet.dcs.napier.ac.uk/evoweb/resources/nutshell/

    http://www.ai-junkie.com/gat1.htm

    http://cs.felk.cvut.cz/~xobitko/ga/

    http://www.secs.ex.ac.uk/resource/IT/ga.html (we have extensive

    research interests in GAs and evolutionary computation)

  • 8/8/2019 Robotics 16

    10/13

    Artificial Life (Alife)

    Each cell has 8 possible neighbours.

    1. If a cell is alive, it will survive

    in the next generation if there

    are either two or three

    neighbours also alive.

    2. If a cell is alive, it will die of

    overcrowding if there are morethan three live neighbours.

    3. If a cell is alive, it will die of

    overexposure if there are fewer

    than two of its neighbours

    alive.

    4. If a cell is dead, it will remain

    dead unless exactly three of its

    neighbours are alive, in which

    case it will be born in the next

    generation.

    Life occurs on a virtual checkerboard. The

    squares are cells. They are alive or dead.

  • 8/8/2019 Robotics 16

    11/13

    Introduction to emergentism

    An example of large scale events emerging from events

    at a lower level.

    The five live cells (shaded circles) are surrounded by

    dead cells (open circles) on a virtual checkerboard, where

    each cell has 8 neighbours. Following the rules

    repeatedly, the initial configuration evolves over 4 steps to

    produce a configuration where ...?

  • 8/8/2019 Robotics 16

    12/13

    Exercise

    Using the rules given previously, show three generations of

    the following grid. Note that a blank cell means a dead

    cell.

  • 8/8/2019 Robotics 16

    13/13

    Solution