Académique Documents
Professionnel Documents
Culture Documents
Why is there power in this metaphor? How can they be used to extend the capabilities of the designer?
John Gero Professor of Design Science University of Sydney Visiting Professor of Design and Computation MIT MIT Design Inquiry
l l
Separation of genetic material (genotype [representation]) from organism (phenotype [design]) Expressing genotype as organism Organism carries genotype and reproduces genotype using genetic processes of crossover and mutation Darwins natural selection uses fitnesses of organisms in their environment to improve the gene pool GA is a simple model of this process
MIT Design Inquiry
Genetic processes
Crossover Points
A Parents
Offspring C
MIT Design Inquiry
Parents
Offspring
Crossover point
MIT Design Inquiry
Parents
Offspring
Crossover point
MIT Design Inquiry
Can map genetic representation onto a computational representation of a design; can map phenotype onto a interpretable view of a design Humans work on single or few designs at a time/ genetics works on a population of designs in parallel Humans can be seen to search design spaces - this is one interpretation of what GAs are doing.
l l l l l l
Guaranteed improvement - Darwinian evolution Large scale search Blind search Fitness can be human evaluation Fitness can change over evolutionary time Can produce complexity Can produce unexpected results
Creativity through genetic engineering Novel designs through extending genetic crossover Novel designs through different fitnesses
l l
Background l genes, genotype, phenotype, fitness Connecting genes to performance in fitness Emergent gene clusters evolved genes
x x x x
good genotypes
bad genotypes
b a rule 7 b a a rule 8 a
design 1
design 2
design 3
good
{1,12,2,8,5,4,4,2,8,5,7} design 4 design 5
good
{1,2,1,8,2,8,5,5,6,6,8,1} design 6
bad
{3,2,2,6,5,8,2,1,4,4,3,1}
good
{6,4,1,2,8,5,4,2,8,5,3,3} design 7 design 8
bad
{3,4,8,2,8,1,6,5,7,3} design 9
neutral
{2,3,2,3,4,3,5,6,5,1,6,2} design 10
good
bad
neutral
bad
Mondrian
Genotype Form In form of a tree Each node has four variables direction of rectangular split (4 values) fraction of the split (15 values) colour of split area (10 values) line width (3 values)
l l l l
l l
l l l
offset between actual and required positions of dissection lines number of lines with correct line width, normalised number of correct colour panels, normalised number of lines assigned, normalised number of unassigned lines, normalised
Flondrians
l
Mondrian painting genetically engineered genes: M-genes Frank Lloyd Wright windows genetically engineered genes in same representation: Fgenes Flondrians are the genetic product of mating M-genes with F-genes
g1 gc g2
p2 pc
C(g1,g2) gc
MIT Design Inquiry
C(p1,p2) pc
P+
p1
p2
Interpolation
Image detection
(a)
(b)
Modelling Interest
Reward
1
HEDONIC VALUE
Hx
0 n1 n2
Nx
NOVELTY
-1
Punish
N=0
N=1
N=2
N=3
N=4
N=5
N=6
N=7
N=8
N=9
N=10
N=11
N=12
N=13
N=14
N=15
N=16
N=17
N=18
N=19