SlideShare a Scribd company logo
1 of 25
Download to read offline
NMA 2014 8th Int. Conf. on Numerical Methods and Applications
On Meme Self-Adaptation in Spatially-Structured
Multimemetic Algorithms
Rafael Nogueras Carlos Cotta
Departamento de Lenguajes y Ciencias de la Computación
Universidad de Málaga, Spain
NMA 2014, Borovets, 20-24 August 2014
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
What are Memes?
Memes are information pieces that constitute units of imitation.
“Examples of memes are tunes, ideas,
catch-phrases, clothes fashions, ways of
making pots or of building arches. Just as
genes propagate themselves in the gene pool
by leaping from body to body via sperms or
eggs, so memes propagate themselves in the
meme pool by leaping from brain to brain via a
process which, in the broad sense, can be
called imitation.”
The Selfish Gene, Richard Dawkins, 1976
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
What is a Memetic Algorithm?
Memetic Algorithms
A Memetic Algorithm is a population of agents
that alternate periods of self-improvement
with periods of cooperation, and competition.
Pablo Moscato, 1989
Memes can be implicitly defined be the choice of local-search (i.e.,
self-improvement) method, or can be explicitly described in the
agent.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Multimemetic Algoritms (and Memetic Computing)
The term “multimemetic” was coined by N. Krasnogor and J.
Smith (2001). In an MMA, each agent carries a solution and the
meme(s) to improve it.
Evolution works at these two levels, cf. Moscato (1999).
Memetic Computing
A paradigm that uses the notion of meme(s) as units of
information encoded in computational representations for the
purpose of problem solving.
Ong, Lim, Chen, 2010
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Background and Scope
Some interesting issues in MMAs:
Memes evolve in MMAs alongside with the solutions they
attach to. It is up to the algorithm to (self-adaptively)
discover good fits between genotypes and memes.
Memes can (self-)adapt their complexity during the run.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Background and Scope
Dynamics of meme propagation is more complex than genetic
counterparts.
Genes represent solutions objectively measurable via the
fitness function.
Memes are indirectly evaluated by their effect on solutions.
Previous research by Nogueras and Cotta (2013) indicates that
population structure is very important to determine the behavior of
the algorithm.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Background and Scope
Selection intensity plays a very important role in allowing
high-potential memes to proliferate.
When meme improvement margins are small, average memes can
hitchhike their way to the final stages of the evolution.
Spatial structure induces longer takeover times, mitigating the
hitchhiking effect.
We consider a spatially-structured MMA whose memes are
rewriting rules of fixed/self-adapted length.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Meme Adaptation
Spatial Structure
Meme Representation
Memes are pattern-based rewriting rules [C → A], where:
C, A ∈ Σr with Σ = {0, 1, #} and r ∈ N.
‘#’ represents a wildcard symbol.
A meme [C → A] can be applied on a genotypic string
G = g1 · · · gn, iff
∃i : gi · · · gi+r−1 ' c1 · · · cr
where C = c1 · · · cr . Its action is to implant A in G, i.e.
gi · · · gi+r−1 ← a1 · · · ar
where A = a1 · · · ar .
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Meme Adaptation
Spatial Structure
Meme Application
Example
Let G = 00010011, and let a meme be 01# → 1#0. A possible
application of the meme could be the following:
00
C
z}|{
010 011
meme
−
−
−
−
−
−
−
→ 00 110
|{z}
A
011
Another potential application site would be i = 6, resulting in
00010
C
z}|{
011
meme
−
−
−
−
−
−
−
→ 00010 110
|{z}
A
Genotype scanning is randomized, and at most w sites are used,
keeping the best result.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Meme Adaptation
Spatial Structure
Meme Application
Pseudocode
Algorithm 1: Meme Application
for i ∈ [1 · · · w] do
// w = maximal number of meme applications
repeat
p ← SelectPos(`) // Pick different position
until Gp···p+|C| ' C;
G0
p···p+|A| ← A // Application of meme on genotype
Ni ← G0 // Neighboring genotype is updated
end
if max(f (Ni )) > f (G) then
G ← arg max(f (Ni )) // Individual is updated
end
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Meme Adaptation
Spatial Structure
Spatial Structure
Definition
The spatial structure is characterized using a Boolean µ × µ matrix
S, Sij = true iff the individual at the i-th site can interact with the
individual at the j-th site.
We consider interaction matrices induced by a particular spatial
arrangement of individual sites in a toroidal grid: let µ = a × b,
where:
each site i can then be represented by a pair of coordinates
(ix , iy ) ∈ {1, · · · , a} × {1, · · · , b}.
let d : ({1, · · · , a} × {1, · · · , b})2
→ N be a distance measure
between sites.
Given a certain neighborhood radius ρ, Sij ⇔ (d(i, j) 6 ρ).
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Meme Adaptation
Spatial Structure
Spatial Structure
Neighborhoods
Two sites can interact if they are within a certain distance
threshold.
Different spatial structures depend upon distance measures:
Panmixia: d(·, ·) = 0.
Moore neighborhood:
d((ix , iy ), (jx , jy )) = max(|ix − jx |, |iy − jy |).
von Neumann neighborhood:
d((ix , iy ), (jx , jy )) = |ix − jx | + |iy − jy |.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Meme Adaptation
Spatial Structure
Spatial Structure
Example
Illustration of the different neighborhoods considered.
0 1 2 3 4 5 6 7 8
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6 7 8
0
1
2
3
4
5
6
7
8
0 1 2 3 4 5 6 7 8
0
1
2
3
4
5
6
7
8
The black cell indicates an arbitrary individual and the grey cells
denote its neighbors (ρ = 1).
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Problems Considered
TRAP
A basic 4-bit trap is defined as
ftrap(b1 · · · b4) =
(
0.6 − 0.2 · u(b1 · · · b4) if u(b1 · · · b4) < 4
1 if u(b1 · · · b4) = 4
where u(s1 · · · si ) =
P
j sj is the unitation.
A higher-order problem can be built by concatenating k such traps,
and defining the fitness of a 4k-bit string as the sum of the fitness
contribution of each block.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Problems Considered
HIFF
HIFF is defined for binary strings of 2k bits by using two functions :
f : {0, 1, •} → {0, 1} t : {0, 1, •} → {0, 1, •}
where ‘•’ denotes a null value. These are defined as:
f (a, b) =
(
1 a = b 6= •
0 otherwise
t(a, b) =
(
a a = b
• otherwise
These two functions are combined as follows:
HIFFk(b1 · · · bn) =
n/2
X
i=1
f (b2i−1, b2i ) + 2 · HIFFk−1(b0
1, · · · , b0
n/2)
where b0
i = t(b2i−1, b2i ) and HIFF0(·) = 1.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Configuration
General
These are the global parameters for MMAs:
µ = 100 individuals.
Generational with binary tournament for parent selection.
One-point crossover (pX = 1.0).
Bit-flip mutation (pM = 1/`).
Local-search and replacement of the worst parent.
Spatially structured MMAs consider a 10 × 10 grid and a
neighborhood radius ρ = 1.
25,000 evaluations per execution.
20 runs are performed for each problem and algorithm.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Configuration
Specific
And these are the specific parameters for meme configuration:
Offspring inherit the meme of the best parent.
Meme lengths bounded by lmin = 3 and lmax = 9.
pr = 1/lmax for length self-adaptation.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Numerical Results
Approximation to the Optimum
0.6
0.7
0.8
0.9
1
TRAP HIFF
approximation
to
optimum
panmictic
Moore
von Neumann
Quade test p-value = 1.97 · 10−3.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Numerical Results
Number of Times the Optimum is Found
2 4 6 8
TRAP
HIFF
times optimum found
(a) Panmictic
5 6 7 8 9 10 11
TRAP
HIFF
times optimum found
(b) Moore
6 8 10 12 14
TRAP
HIFF
times optimum found
(c) von Neumann
Quade test p-value = 6.40 · 10−4.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Diversity
Evolution of diversity with von Neumann topology on the TRAP
function.
0.5 1 1.5 2 2.5
x 10
4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
evaluations
entropy
r3
r6
r9
r39
(a) Genetic diversity
0.5 1 1.5 2 2.5
x 10
4
0
0.5
1
1.5
2
2.5
3
3.5
4
evaluations
entropy
r3
r6
r9
r39
(b) Memetic diversity
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Self-Adaptation
Meme Length
Evolution of meme lengths for different topologies and the two
problems considered:
0.5 1 1.5 2 2.5
x 10
4
5
5.2
5.4
5.6
5.8
6
6.2
6.4
6.6
6.8
evaluations
meme
lengths
PANM
Moore
von Neumann
(a) TRAP
0.5 1 1.5 2 2.5
x 10
4
4.8
5
5.2
5.4
5.6
5.8
6
6.2
evaluations
meme
lengths
PANM
Moore
von Neumann
(b) HIFF
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Self-Adaptation
Meme Success Ratio
Percentage of meme applications resulting in an improvement of
the different MMAs on TRAP function:
0.5 1 1.5 2 2.5
x 10
4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
evaluations
improvement
rate
r3
r6
r9
r39
(a) Panmictic
0.5 1 1.5 2 2.5
x 10
4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
evaluations
improvement
rate
r3
r6
r9
r39
(b) Moore
0.5 1 1.5 2 2.5
x 10
4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
evaluations
improvement
rate
r3
r6
r9
r39
(c) von Neumann
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Benchmark and Settings
Results
Self-Adaptation Impact
For fixed-length memes the intermediate value r = 6 offers the
best tradeoff.
For self-adaptive meme length:
Average meme lengths oscillate around values close to 6.
Fully self-adaptive MMA discover this area of memetic
interest.
The role of memes is self-adaptive too, it turns from
exploratory to exploitative.
Success rates follow an upwards trend and the values for
MMA3−9 are normally superior to the remaining MMAs.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Conclusions
Self-adaptation mechanisms alleviating one of the major issues in
memetic algorithms → Parameterization.
Experimental results indicate that self-adaptation of meme lengths
can be beneficial, mainly in panmictic populations, where
suboptimal parameterization is less robust.
Self-Adaptation of meme lengths seems an adequate strategy, since
it does not penalize performance and saves configuration time.
Future work:
further experimentation on other problems,
use of other population structures,
analyze the scalability of the approach.
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
Introduction
Model Description
Experimental Analysis
Conclusions
Thank You!
AnySelf Project
Please find us in Facebook
http://facebook.com/AnySelfProject
and in Twitter
@anyselfproject
R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs

More Related Content

What's hot

Monte Carlo Statistical Methods
Monte Carlo Statistical MethodsMonte Carlo Statistical Methods
Monte Carlo Statistical MethodsChristian Robert
 
P-Systems for approximating NP-Complete optimization problems
P-Systems for approximating NP-Complete optimization problemsP-Systems for approximating NP-Complete optimization problems
P-Systems for approximating NP-Complete optimization problemsFrancesco Corucci
 
Supervised Planetary Unmixing with Optimal Transport
Supervised Planetary Unmixing with Optimal TransportSupervised Planetary Unmixing with Optimal Transport
Supervised Planetary Unmixing with Optimal TransportSina Nakhostin
 
Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Alexander Litvinenko
 
(DL輪読)Matching Networks for One Shot Learning
(DL輪読)Matching Networks for One Shot Learning(DL輪読)Matching Networks for One Shot Learning
(DL輪読)Matching Networks for One Shot LearningMasahiro Suzuki
 
Markov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing themMarkov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing themPierre Jacob
 
Metropolis-Hastings MCMC Short Tutorial
Metropolis-Hastings MCMC Short TutorialMetropolis-Hastings MCMC Short Tutorial
Metropolis-Hastings MCMC Short TutorialRalph Schlosser
 
Random Features Strengthen Graph Neural Networks
Random Features Strengthen Graph Neural NetworksRandom Features Strengthen Graph Neural Networks
Random Features Strengthen Graph Neural Networksjoisino
 
Introduction to MCMC methods
Introduction to MCMC methodsIntroduction to MCMC methods
Introduction to MCMC methodsChristian Robert
 
Fast Unbalanced Optimal Transport on a Tree
Fast Unbalanced Optimal Transport on a TreeFast Unbalanced Optimal Transport on a Tree
Fast Unbalanced Optimal Transport on a Treejoisino
 
Information Content of Complex Networks
Information Content of Complex NetworksInformation Content of Complex Networks
Information Content of Complex NetworksHector Zenil
 
Monash University short course, part I
Monash University short course, part IMonash University short course, part I
Monash University short course, part IChristian Robert
 
Mcmc & lkd free I
Mcmc & lkd free IMcmc & lkd free I
Mcmc & lkd free IDeb Roy
 
Peridynamics MECOM 2012 presentation
Peridynamics MECOM 2012 presentationPeridynamics MECOM 2012 presentation
Peridynamics MECOM 2012 presentationDiego Tramontina
 
Neural Networks: Radial Bases Functions (RBF)
Neural Networks: Radial Bases Functions (RBF)Neural Networks: Radial Bases Functions (RBF)
Neural Networks: Radial Bases Functions (RBF)Mostafa G. M. Mostafa
 
Kernel methods for data integration in systems biology
Kernel methods for data integration in systems biology Kernel methods for data integration in systems biology
Kernel methods for data integration in systems biology tuxette
 

What's hot (20)

Monte Carlo Statistical Methods
Monte Carlo Statistical MethodsMonte Carlo Statistical Methods
Monte Carlo Statistical Methods
 
P-Systems for approximating NP-Complete optimization problems
P-Systems for approximating NP-Complete optimization problemsP-Systems for approximating NP-Complete optimization problems
P-Systems for approximating NP-Complete optimization problems
 
Supervised Planetary Unmixing with Optimal Transport
Supervised Planetary Unmixing with Optimal TransportSupervised Planetary Unmixing with Optimal Transport
Supervised Planetary Unmixing with Optimal Transport
 
Analysis_molf
Analysis_molfAnalysis_molf
Analysis_molf
 
Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...
 
Dycops2019
Dycops2019 Dycops2019
Dycops2019
 
(DL輪読)Matching Networks for One Shot Learning
(DL輪読)Matching Networks for One Shot Learning(DL輪読)Matching Networks for One Shot Learning
(DL輪読)Matching Networks for One Shot Learning
 
Markov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing themMarkov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing them
 
Metropolis-Hastings MCMC Short Tutorial
Metropolis-Hastings MCMC Short TutorialMetropolis-Hastings MCMC Short Tutorial
Metropolis-Hastings MCMC Short Tutorial
 
Random Features Strengthen Graph Neural Networks
Random Features Strengthen Graph Neural NetworksRandom Features Strengthen Graph Neural Networks
Random Features Strengthen Graph Neural Networks
 
Introduction to MCMC methods
Introduction to MCMC methodsIntroduction to MCMC methods
Introduction to MCMC methods
 
Fast Unbalanced Optimal Transport on a Tree
Fast Unbalanced Optimal Transport on a TreeFast Unbalanced Optimal Transport on a Tree
Fast Unbalanced Optimal Transport on a Tree
 
Information Content of Complex Networks
Information Content of Complex NetworksInformation Content of Complex Networks
Information Content of Complex Networks
 
Monash University short course, part I
Monash University short course, part IMonash University short course, part I
Monash University short course, part I
 
Mcmc & lkd free I
Mcmc & lkd free IMcmc & lkd free I
Mcmc & lkd free I
 
Talk 5
Talk 5Talk 5
Talk 5
 
Peridynamics MECOM 2012 presentation
Peridynamics MECOM 2012 presentationPeridynamics MECOM 2012 presentation
Peridynamics MECOM 2012 presentation
 
Statistical Physics Studies of Machine Learning Problems by Lenka Zdeborova, ...
Statistical Physics Studies of Machine Learning Problems by Lenka Zdeborova, ...Statistical Physics Studies of Machine Learning Problems by Lenka Zdeborova, ...
Statistical Physics Studies of Machine Learning Problems by Lenka Zdeborova, ...
 
Neural Networks: Radial Bases Functions (RBF)
Neural Networks: Radial Bases Functions (RBF)Neural Networks: Radial Bases Functions (RBF)
Neural Networks: Radial Bases Functions (RBF)
 
Kernel methods for data integration in systems biology
Kernel methods for data integration in systems biology Kernel methods for data integration in systems biology
Kernel methods for data integration in systems biology
 

Similar to On Meme Self-Adaptation in Spatially-Structured Multimemetic Algorithms

An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...
An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...
An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...Rafael Nogueras
 
Person re-identification, PhD Day 2011
Person re-identification, PhD Day 2011Person re-identification, PhD Day 2011
Person re-identification, PhD Day 2011Riccardo Satta
 
Multi objective predictive control a solution using metaheuristics
Multi objective predictive control  a solution using metaheuristicsMulti objective predictive control  a solution using metaheuristics
Multi objective predictive control a solution using metaheuristicsijcsit
 
Microstrip coupler design using bat
Microstrip coupler design using batMicrostrip coupler design using bat
Microstrip coupler design using batijaia
 
Medical diagnosis classification
Medical diagnosis classificationMedical diagnosis classification
Medical diagnosis classificationcsandit
 
MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...
MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...
MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...cscpconf
 
Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...
Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...
Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...ArchiLab 7
 
An Efficient Genetic Algorithm for Solving Knapsack Problem.pdf
An Efficient Genetic Algorithm for Solving Knapsack Problem.pdfAn Efficient Genetic Algorithm for Solving Knapsack Problem.pdf
An Efficient Genetic Algorithm for Solving Knapsack Problem.pdfNancy Ideker
 
Artificial Intelligence Applications in Petroleum Engineering - Part I
Artificial Intelligence Applications in Petroleum Engineering - Part IArtificial Intelligence Applications in Petroleum Engineering - Part I
Artificial Intelligence Applications in Petroleum Engineering - Part IRamez Abdalla, M.Sc
 
nonlinear_rmt.pdf
nonlinear_rmt.pdfnonlinear_rmt.pdf
nonlinear_rmt.pdfGieTe
 
Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010
Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010
Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010Darin Hitchings, Ph.D.
 
An Information-Theoretic Approach for Clonal Selection Algorithms
An Information-Theoretic Approach for Clonal Selection AlgorithmsAn Information-Theoretic Approach for Clonal Selection Algorithms
An Information-Theoretic Approach for Clonal Selection AlgorithmsMario Pavone
 
Self-Assembling Hyper-heuristics: a proof of concept
Self-Assembling Hyper-heuristics: a proof of conceptSelf-Assembling Hyper-heuristics: a proof of concept
Self-Assembling Hyper-heuristics: a proof of conceptGerman Terrazas
 
Self scale estimation of the tracking window merged with adaptive particle fi...
Self scale estimation of the tracking window merged with adaptive particle fi...Self scale estimation of the tracking window merged with adaptive particle fi...
Self scale estimation of the tracking window merged with adaptive particle fi...IJECEIAES
 
M ESH S IMPLIFICATION V IA A V OLUME C OST M EASURE
M ESH S IMPLIFICATION V IA A V OLUME C OST M EASUREM ESH S IMPLIFICATION V IA A V OLUME C OST M EASURE
M ESH S IMPLIFICATION V IA A V OLUME C OST M EASUREijcga
 

Similar to On Meme Self-Adaptation in Spatially-Structured Multimemetic Algorithms (20)

An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...
An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...
An Analysis of a Selecto-Lamarckian Model of Multimemetic Algorithms with Dyn...
 
Person re-identification, PhD Day 2011
Person re-identification, PhD Day 2011Person re-identification, PhD Day 2011
Person re-identification, PhD Day 2011
 
Multi objective predictive control a solution using metaheuristics
Multi objective predictive control  a solution using metaheuristicsMulti objective predictive control  a solution using metaheuristics
Multi objective predictive control a solution using metaheuristics
 
Microstrip coupler design using bat
Microstrip coupler design using batMicrostrip coupler design using bat
Microstrip coupler design using bat
 
Medical diagnosis classification
Medical diagnosis classificationMedical diagnosis classification
Medical diagnosis classification
 
MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...
MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...
MEDICAL DIAGNOSIS CLASSIFICATION USING MIGRATION BASED DIFFERENTIAL EVOLUTION...
 
Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...
Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...
Ramos, almeida: artificial ant colonies in digital image habitats – a mass be...
 
Presentation v3.2
Presentation v3.2Presentation v3.2
Presentation v3.2
 
Presentation v3.2
Presentation v3.2Presentation v3.2
Presentation v3.2
 
An Efficient Genetic Algorithm for Solving Knapsack Problem.pdf
An Efficient Genetic Algorithm for Solving Knapsack Problem.pdfAn Efficient Genetic Algorithm for Solving Knapsack Problem.pdf
An Efficient Genetic Algorithm for Solving Knapsack Problem.pdf
 
Artificial Intelligence Applications in Petroleum Engineering - Part I
Artificial Intelligence Applications in Petroleum Engineering - Part IArtificial Intelligence Applications in Petroleum Engineering - Part I
Artificial Intelligence Applications in Petroleum Engineering - Part I
 
nonlinear_rmt.pdf
nonlinear_rmt.pdfnonlinear_rmt.pdf
nonlinear_rmt.pdf
 
Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010
Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010
Receding Horizon Stochastic Control Algorithms for Sensor Management ACC 2010
 
An Information-Theoretic Approach for Clonal Selection Algorithms
An Information-Theoretic Approach for Clonal Selection AlgorithmsAn Information-Theoretic Approach for Clonal Selection Algorithms
An Information-Theoretic Approach for Clonal Selection Algorithms
 
Self-Assembling Hyper-heuristics: a proof of concept
Self-Assembling Hyper-heuristics: a proof of conceptSelf-Assembling Hyper-heuristics: a proof of concept
Self-Assembling Hyper-heuristics: a proof of concept
 
Self scale estimation of the tracking window merged with adaptive particle fi...
Self scale estimation of the tracking window merged with adaptive particle fi...Self scale estimation of the tracking window merged with adaptive particle fi...
Self scale estimation of the tracking window merged with adaptive particle fi...
 
M ESH S IMPLIFICATION V IA A V OLUME C OST M EASURE
M ESH S IMPLIFICATION V IA A V OLUME C OST M EASUREM ESH S IMPLIFICATION V IA A V OLUME C OST M EASURE
M ESH S IMPLIFICATION V IA A V OLUME C OST M EASURE
 
September 9, Deliberative Algorithms I
September 9, Deliberative Algorithms ISeptember 9, Deliberative Algorithms I
September 9, Deliberative Algorithms I
 
ICPR Workshop ETCHB 2010
ICPR Workshop ETCHB 2010ICPR Workshop ETCHB 2010
ICPR Workshop ETCHB 2010
 
CoopLoc Technical Presentation
CoopLoc Technical PresentationCoopLoc Technical Presentation
CoopLoc Technical Presentation
 

More from Rafael Nogueras

Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...
Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...
Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...Rafael Nogueras
 
A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...
A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...
A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...Rafael Nogueras
 
Evaluating Island-based EAs on Unstable Networks with Complex Failure Patterns
Evaluating Island-based EAs on Unstable Networks with Complex Failure PatternsEvaluating Island-based EAs on Unstable Networks with Complex Failure Patterns
Evaluating Island-based EAs on Unstable Networks with Complex Failure PatternsRafael Nogueras
 
A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...
A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...
A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...Rafael Nogueras
 
A Study on Multimemetic Estimation of Distribution Algorithms
A Study on Multimemetic Estimation of Distribution AlgorithmsA Study on Multimemetic Estimation of Distribution Algorithms
A Study on Multimemetic Estimation of Distribution AlgorithmsRafael Nogueras
 
An Analysis of Migration Strategies in Island-Based Multimemetic Algorithms
An Analysis of Migration Strategies in Island-Based Multimemetic AlgorithmsAn Analysis of Migration Strategies in Island-Based Multimemetic Algorithms
An Analysis of Migration Strategies in Island-Based Multimemetic AlgorithmsRafael Nogueras
 
Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...
Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...
Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...Rafael Nogueras
 
On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...
On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...
On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...Rafael Nogueras
 
Análisis de llamadas telefónicas con Teoría de Grafos y R
Análisis de llamadas telefónicas con Teoría de Grafos y RAnálisis de llamadas telefónicas con Teoría de Grafos y R
Análisis de llamadas telefónicas con Teoría de Grafos y RRafael Nogueras
 

More from Rafael Nogueras (9)

Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...
Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...
Analyzing Resilience to Computational Glitches in Island-based Evolutionary A...
 
A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...
A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...
A Performance Analysis of Self-* Evolutionary Algorithms on Networks with Cor...
 
Evaluating Island-based EAs on Unstable Networks with Complex Failure Patterns
Evaluating Island-based EAs on Unstable Networks with Complex Failure PatternsEvaluating Island-based EAs on Unstable Networks with Complex Failure Patterns
Evaluating Island-based EAs on Unstable Networks with Complex Failure Patterns
 
A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...
A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...
A Study of the Performance of Self-* Memetic Algorithms on Heterogeneous Ephe...
 
A Study on Multimemetic Estimation of Distribution Algorithms
A Study on Multimemetic Estimation of Distribution AlgorithmsA Study on Multimemetic Estimation of Distribution Algorithms
A Study on Multimemetic Estimation of Distribution Algorithms
 
An Analysis of Migration Strategies in Island-Based Multimemetic Algorithms
An Analysis of Migration Strategies in Island-Based Multimemetic AlgorithmsAn Analysis of Migration Strategies in Island-Based Multimemetic Algorithms
An Analysis of Migration Strategies in Island-Based Multimemetic Algorithms
 
Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...
Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...
Introducción de Sesgo Heurístico en Algoritmos Genéticos Mediante Operadores ...
 
On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...
On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...
On Partitioned Fitness Distributions of Genetic Operators for Predicting GA P...
 
Análisis de llamadas telefónicas con Teoría de Grafos y R
Análisis de llamadas telefónicas con Teoría de Grafos y RAnálisis de llamadas telefónicas con Teoría de Grafos y R
Análisis de llamadas telefónicas con Teoría de Grafos y R
 

Recently uploaded

The dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxThe dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxEran Akiva Sinbar
 
Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.aasikanpl
 
Solution chemistry, Moral and Normal solutions
Solution chemistry, Moral and Normal solutionsSolution chemistry, Moral and Normal solutions
Solution chemistry, Moral and Normal solutionsHajira Mahmood
 
Twin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptxTwin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptxEran Akiva Sinbar
 
Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2John Carlo Rollon
 
OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024innovationoecd
 
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝soniya singh
 
Pests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPirithiRaju
 
Forest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are importantForest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are importantadityabhardwaj282
 
RESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptx
RESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptxRESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptx
RESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptxFarihaAbdulRasheed
 
GenBio2 - Lesson 1 - Introduction to Genetics.pptx
GenBio2 - Lesson 1 - Introduction to Genetics.pptxGenBio2 - Lesson 1 - Introduction to Genetics.pptx
GenBio2 - Lesson 1 - Introduction to Genetics.pptxBerniceCayabyab1
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxpriyankatabhane
 
Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...Nistarini College, Purulia (W.B) India
 
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)riyaescorts54
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Patrick Diehl
 
zoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistanzoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistanzohaibmir069
 
Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫qfactory1
 
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxMicrophone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxpriyankatabhane
 

Recently uploaded (20)

The dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptxThe dark energy paradox leads to a new structure of spacetime.pptx
The dark energy paradox leads to a new structure of spacetime.pptx
 
Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Munirka Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
 
Solution chemistry, Moral and Normal solutions
Solution chemistry, Moral and Normal solutionsSolution chemistry, Moral and Normal solutions
Solution chemistry, Moral and Normal solutions
 
Twin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptxTwin's paradox experiment is a meassurement of the extra dimensions.pptx
Twin's paradox experiment is a meassurement of the extra dimensions.pptx
 
Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2Evidences of Evolution General Biology 2
Evidences of Evolution General Biology 2
 
OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024OECD bibliometric indicators: Selected highlights, April 2024
OECD bibliometric indicators: Selected highlights, April 2024
 
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
Call Girls in Munirka Delhi 💯Call Us 🔝8264348440🔝
 
Pests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdfPests of castor_Binomics_Identification_Dr.UPR.pdf
Pests of castor_Binomics_Identification_Dr.UPR.pdf
 
Forest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are importantForest laws, Indian forest laws, why they are important
Forest laws, Indian forest laws, why they are important
 
RESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptx
RESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptxRESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptx
RESPIRATORY ADAPTATIONS TO HYPOXIA IN HUMNAS.pptx
 
GenBio2 - Lesson 1 - Introduction to Genetics.pptx
GenBio2 - Lesson 1 - Introduction to Genetics.pptxGenBio2 - Lesson 1 - Introduction to Genetics.pptx
GenBio2 - Lesson 1 - Introduction to Genetics.pptx
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptx
 
Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...Bentham & Hooker's Classification. along with the merits and demerits of the ...
Bentham & Hooker's Classification. along with the merits and demerits of the ...
 
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
(9818099198) Call Girls In Noida Sector 14 (NOIDA ESCORTS)
 
Engler and Prantl system of classification in plant taxonomy
Engler and Prantl system of classification in plant taxonomyEngler and Prantl system of classification in plant taxonomy
Engler and Prantl system of classification in plant taxonomy
 
Volatile Oils Pharmacognosy And Phytochemistry -I
Volatile Oils Pharmacognosy And Phytochemistry -IVolatile Oils Pharmacognosy And Phytochemistry -I
Volatile Oils Pharmacognosy And Phytochemistry -I
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?
 
zoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistanzoogeography of pakistan.pptx fauna of Pakistan
zoogeography of pakistan.pptx fauna of Pakistan
 
Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫Manassas R - Parkside Middle School 🌎🏫
Manassas R - Parkside Middle School 🌎🏫
 
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptxMicrophone- characteristics,carbon microphone, dynamic microphone.pptx
Microphone- characteristics,carbon microphone, dynamic microphone.pptx
 

On Meme Self-Adaptation in Spatially-Structured Multimemetic Algorithms

  • 1. NMA 2014 8th Int. Conf. on Numerical Methods and Applications On Meme Self-Adaptation in Spatially-Structured Multimemetic Algorithms Rafael Nogueras Carlos Cotta Departamento de Lenguajes y Ciencias de la Computación Universidad de Málaga, Spain NMA 2014, Borovets, 20-24 August 2014 R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 2. Introduction Model Description Experimental Analysis Conclusions What are Memes? Memes are information pieces that constitute units of imitation. “Examples of memes are tunes, ideas, catch-phrases, clothes fashions, ways of making pots or of building arches. Just as genes propagate themselves in the gene pool by leaping from body to body via sperms or eggs, so memes propagate themselves in the meme pool by leaping from brain to brain via a process which, in the broad sense, can be called imitation.” The Selfish Gene, Richard Dawkins, 1976 R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 3. Introduction Model Description Experimental Analysis Conclusions What is a Memetic Algorithm? Memetic Algorithms A Memetic Algorithm is a population of agents that alternate periods of self-improvement with periods of cooperation, and competition. Pablo Moscato, 1989 Memes can be implicitly defined be the choice of local-search (i.e., self-improvement) method, or can be explicitly described in the agent. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 4. Introduction Model Description Experimental Analysis Conclusions Multimemetic Algoritms (and Memetic Computing) The term “multimemetic” was coined by N. Krasnogor and J. Smith (2001). In an MMA, each agent carries a solution and the meme(s) to improve it. Evolution works at these two levels, cf. Moscato (1999). Memetic Computing A paradigm that uses the notion of meme(s) as units of information encoded in computational representations for the purpose of problem solving. Ong, Lim, Chen, 2010 R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 5. Introduction Model Description Experimental Analysis Conclusions Background and Scope Some interesting issues in MMAs: Memes evolve in MMAs alongside with the solutions they attach to. It is up to the algorithm to (self-adaptively) discover good fits between genotypes and memes. Memes can (self-)adapt their complexity during the run. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 6. Introduction Model Description Experimental Analysis Conclusions Background and Scope Dynamics of meme propagation is more complex than genetic counterparts. Genes represent solutions objectively measurable via the fitness function. Memes are indirectly evaluated by their effect on solutions. Previous research by Nogueras and Cotta (2013) indicates that population structure is very important to determine the behavior of the algorithm. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 7. Introduction Model Description Experimental Analysis Conclusions Background and Scope Selection intensity plays a very important role in allowing high-potential memes to proliferate. When meme improvement margins are small, average memes can hitchhike their way to the final stages of the evolution. Spatial structure induces longer takeover times, mitigating the hitchhiking effect. We consider a spatially-structured MMA whose memes are rewriting rules of fixed/self-adapted length. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 8. Introduction Model Description Experimental Analysis Conclusions Meme Adaptation Spatial Structure Meme Representation Memes are pattern-based rewriting rules [C → A], where: C, A ∈ Σr with Σ = {0, 1, #} and r ∈ N. ‘#’ represents a wildcard symbol. A meme [C → A] can be applied on a genotypic string G = g1 · · · gn, iff ∃i : gi · · · gi+r−1 ' c1 · · · cr where C = c1 · · · cr . Its action is to implant A in G, i.e. gi · · · gi+r−1 ← a1 · · · ar where A = a1 · · · ar . R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 9. Introduction Model Description Experimental Analysis Conclusions Meme Adaptation Spatial Structure Meme Application Example Let G = 00010011, and let a meme be 01# → 1#0. A possible application of the meme could be the following: 00 C z}|{ 010 011 meme − − − − − − − → 00 110 |{z} A 011 Another potential application site would be i = 6, resulting in 00010 C z}|{ 011 meme − − − − − − − → 00010 110 |{z} A Genotype scanning is randomized, and at most w sites are used, keeping the best result. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 10. Introduction Model Description Experimental Analysis Conclusions Meme Adaptation Spatial Structure Meme Application Pseudocode Algorithm 1: Meme Application for i ∈ [1 · · · w] do // w = maximal number of meme applications repeat p ← SelectPos(`) // Pick different position until Gp···p+|C| ' C; G0 p···p+|A| ← A // Application of meme on genotype Ni ← G0 // Neighboring genotype is updated end if max(f (Ni )) > f (G) then G ← arg max(f (Ni )) // Individual is updated end R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 11. Introduction Model Description Experimental Analysis Conclusions Meme Adaptation Spatial Structure Spatial Structure Definition The spatial structure is characterized using a Boolean µ × µ matrix S, Sij = true iff the individual at the i-th site can interact with the individual at the j-th site. We consider interaction matrices induced by a particular spatial arrangement of individual sites in a toroidal grid: let µ = a × b, where: each site i can then be represented by a pair of coordinates (ix , iy ) ∈ {1, · · · , a} × {1, · · · , b}. let d : ({1, · · · , a} × {1, · · · , b})2 → N be a distance measure between sites. Given a certain neighborhood radius ρ, Sij ⇔ (d(i, j) 6 ρ). R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 12. Introduction Model Description Experimental Analysis Conclusions Meme Adaptation Spatial Structure Spatial Structure Neighborhoods Two sites can interact if they are within a certain distance threshold. Different spatial structures depend upon distance measures: Panmixia: d(·, ·) = 0. Moore neighborhood: d((ix , iy ), (jx , jy )) = max(|ix − jx |, |iy − jy |). von Neumann neighborhood: d((ix , iy ), (jx , jy )) = |ix − jx | + |iy − jy |. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 13. Introduction Model Description Experimental Analysis Conclusions Meme Adaptation Spatial Structure Spatial Structure Example Illustration of the different neighborhoods considered. 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 The black cell indicates an arbitrary individual and the grey cells denote its neighbors (ρ = 1). R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 14. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Problems Considered TRAP A basic 4-bit trap is defined as ftrap(b1 · · · b4) = ( 0.6 − 0.2 · u(b1 · · · b4) if u(b1 · · · b4) < 4 1 if u(b1 · · · b4) = 4 where u(s1 · · · si ) = P j sj is the unitation. A higher-order problem can be built by concatenating k such traps, and defining the fitness of a 4k-bit string as the sum of the fitness contribution of each block. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 15. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Problems Considered HIFF HIFF is defined for binary strings of 2k bits by using two functions : f : {0, 1, •} → {0, 1} t : {0, 1, •} → {0, 1, •} where ‘•’ denotes a null value. These are defined as: f (a, b) = ( 1 a = b 6= • 0 otherwise t(a, b) = ( a a = b • otherwise These two functions are combined as follows: HIFFk(b1 · · · bn) = n/2 X i=1 f (b2i−1, b2i ) + 2 · HIFFk−1(b0 1, · · · , b0 n/2) where b0 i = t(b2i−1, b2i ) and HIFF0(·) = 1. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 16. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Configuration General These are the global parameters for MMAs: µ = 100 individuals. Generational with binary tournament for parent selection. One-point crossover (pX = 1.0). Bit-flip mutation (pM = 1/`). Local-search and replacement of the worst parent. Spatially structured MMAs consider a 10 × 10 grid and a neighborhood radius ρ = 1. 25,000 evaluations per execution. 20 runs are performed for each problem and algorithm. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 17. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Configuration Specific And these are the specific parameters for meme configuration: Offspring inherit the meme of the best parent. Meme lengths bounded by lmin = 3 and lmax = 9. pr = 1/lmax for length self-adaptation. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 18. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Numerical Results Approximation to the Optimum 0.6 0.7 0.8 0.9 1 TRAP HIFF approximation to optimum panmictic Moore von Neumann Quade test p-value = 1.97 · 10−3. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 19. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Numerical Results Number of Times the Optimum is Found 2 4 6 8 TRAP HIFF times optimum found (a) Panmictic 5 6 7 8 9 10 11 TRAP HIFF times optimum found (b) Moore 6 8 10 12 14 TRAP HIFF times optimum found (c) von Neumann Quade test p-value = 6.40 · 10−4. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 20. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Diversity Evolution of diversity with von Neumann topology on the TRAP function. 0.5 1 1.5 2 2.5 x 10 4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 evaluations entropy r3 r6 r9 r39 (a) Genetic diversity 0.5 1 1.5 2 2.5 x 10 4 0 0.5 1 1.5 2 2.5 3 3.5 4 evaluations entropy r3 r6 r9 r39 (b) Memetic diversity R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 21. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Self-Adaptation Meme Length Evolution of meme lengths for different topologies and the two problems considered: 0.5 1 1.5 2 2.5 x 10 4 5 5.2 5.4 5.6 5.8 6 6.2 6.4 6.6 6.8 evaluations meme lengths PANM Moore von Neumann (a) TRAP 0.5 1 1.5 2 2.5 x 10 4 4.8 5 5.2 5.4 5.6 5.8 6 6.2 evaluations meme lengths PANM Moore von Neumann (b) HIFF R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 22. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Self-Adaptation Meme Success Ratio Percentage of meme applications resulting in an improvement of the different MMAs on TRAP function: 0.5 1 1.5 2 2.5 x 10 4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 evaluations improvement rate r3 r6 r9 r39 (a) Panmictic 0.5 1 1.5 2 2.5 x 10 4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 evaluations improvement rate r3 r6 r9 r39 (b) Moore 0.5 1 1.5 2 2.5 x 10 4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 evaluations improvement rate r3 r6 r9 r39 (c) von Neumann R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 23. Introduction Model Description Experimental Analysis Conclusions Benchmark and Settings Results Self-Adaptation Impact For fixed-length memes the intermediate value r = 6 offers the best tradeoff. For self-adaptive meme length: Average meme lengths oscillate around values close to 6. Fully self-adaptive MMA discover this area of memetic interest. The role of memes is self-adaptive too, it turns from exploratory to exploitative. Success rates follow an upwards trend and the values for MMA3−9 are normally superior to the remaining MMAs. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 24. Introduction Model Description Experimental Analysis Conclusions Conclusions Self-adaptation mechanisms alleviating one of the major issues in memetic algorithms → Parameterization. Experimental results indicate that self-adaptation of meme lengths can be beneficial, mainly in panmictic populations, where suboptimal parameterization is less robust. Self-Adaptation of meme lengths seems an adequate strategy, since it does not penalize performance and saves configuration time. Future work: further experimentation on other problems, use of other population structures, analyze the scalability of the approach. R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs
  • 25. Introduction Model Description Experimental Analysis Conclusions Thank You! AnySelf Project Please find us in Facebook http://facebook.com/AnySelfProject and in Twitter @anyselfproject R. Nogueras, C. Cotta Meme Self-Adaptation in Spatially-Structured MMAs