SlideShare a Scribd company logo
1 of 19
Download to read offline
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
. 
. 
Covariance Matrix Adaptation Evolution 
Strategies(CMA-ES) 
Hossein Abedi 
Evolutionary Computation 
Autumn 2014 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 1 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Overview 
1. Introduction 
2. Selection and Recombination 
3. Adaptation of covariance matrix 
4. Step size control 
5. Experiments 
6. Conclusion 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 2 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Introduction 
Idea 
Introduced by Hansen and Ostermeier in 2001 
The idea: 
Figure : Movement toward a minimum through 3 generations 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 3 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Selection and Recombination 
Generating the children 
New points are sampled normally distributed: 
Xi  M + Ni (0,C ), for i=1,..., 
Figure : Different shapes of C as a hyperelipsoid in 2D 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 4 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Selection and Recombination 
Selection and Recombination 
The mean vector M 2 ℜn is calculated Σas the weighted average of the 
best candidate solutions: M= 
 
i=1 wiXi : 
Where: Σ 
i=1 wi = 1 
w1 ⩾ w2 ⩾ ::: ⩾ w  0 
f (X1:) ⩽ f (X2:) ⩽ ::: ⩽ f (X:) 
eff = ( 
jjwjj1 
jjwjj2 
)2 = Σ 1  
i=1 w2 
i 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 5 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Estimating the covariance matrix from scratch 
For the sake of simplicity set (g) = 1 
Estimating distribution within the population: 
C(g+1) 
emp = 1 
1 
Σ 
i=1(X(g+1) 
i 
 1 
 
Σ 
j=1 Xj )(X(g+1) 
i 
 1 
 
Σ 
j=1 Xj )T 
Estimating distribution of sampled steps: 
C(g+1) 
 = 1 
 
Σ 
i=1(X(g+1) 
i 
 M(g))(X(g+1) 
i 
 M(g))T 
Where: 
The sampled steps are: 
X(g+1) 
 M(g) 
i 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 6 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Estimating the covariance matrix 
Estimating distribution of the most successful steps: 
C(g+1) 
 = 1 
 
Σ 
i=1 wi (X(g+1) 
i : 
 M(g))(X(g+1) 
i : 
 M(g))T 
Estimation of Multivariate Normal Algorithm(ENMA): 
C(g+1) 
 = 1 
 
Σ 
i=1(X(g+1) 
i : 
 M(g+1) 
enma )(X(g+1) 
i : 
 M(g+1) 
enma )T 
Where: 
M(g+1) 
enma = 1 
 
Σ 
j=1 X(g+1) 
j : 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 7 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Estimating the covariance matrix 
Comparison: 
Figure : Covariance matrix estimation on f (x1; x2) = Σ2 
i=1 xi 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 8 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Rank  update 
Smaller  means faster but less global search 
To give recent generations a higher weight, consider a leraning rate c 
and the equation below: 
C(g+1) = (1  c)C(g) + c 
1 
C(g+1) 
((g))2 
Where: 
1 
c 
is called the time back horizon 
Figure : Example of exponential smoothing 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 9 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Rank  update 
C(g+1) = (1  c)C(g) + c 
1 
 
Σ 
i=1 wiOP(X(g+1) 
i : 
M(g) 
(g) ) 
Where: 
OP(y) = yyT 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 10 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Rank one update 
Evolution Path (Pc 2 ℜn): sum of consecutive steps: 
M(g+1)M(g) 
(g) + M(g)M(g1) 
(g1) + ::: 
Figure : Evolution path 
N(0; I )y1 + N(0; I )y2 + ::: + N(0; I )yg  N(0; 
Σg 
i=1 yi yT 
i ) 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 11 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Rank one update 
Using exponential smoothing: 
P(g+1) 
c = (1  cc )P(g) 
c + 
√ 
cc(2  cc )eff 
M(g+1)M(g) 
(g) 
Wher√e: 
cc(2  cc )eff is a scaling factor such that :P(g+1) 
c  N(0; C) 
So rank one update with sign is : 
C(g+1) = (1  c1)C(g) + c1OP(P(g+1) 
c ) 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 12 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Adaptation of covariance matrix 
Cumulation 
C(g+1) = (1  c1  c)C(g) + c1(y(g+1) 
c )(P(g+1) 
c )T + 
:::c 
1 
 
Σ 
i=1 wiOP(X(g+1) 
i : 
M(g) 
(g) ) 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 13 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Step size control 
Step size control 
Using the evolution path for adapting the stepsize  
Figure : Different evolution path senarios for 6 consecutive mean vectors 
(g+1) = (g) exp ( c 
d 
( 
jjp(g+1) 
 jj 
EjjN(0;I )jj  1)) 
Where: 
p(g+1) 
 = (1  c)p(g) 
 + 
√ 
c(2  c)eff (C(g))1 
2 M(g+1)M(g) 
(g) 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 14 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Experiments 
Test on seperable and non rotated 
0 200 400 600 800 1000 1200 1400 1600 1800 2000 
14 
12 
10 
8 
6 
4 
2 
0 
0.01*function evauations 
fmin 
CLPSO 
CMA−ES 
Figure : Results on Ackley test function 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 15 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Experiments 
Test on CEC2015(shifted,rotated and non-seperable) 
0 10 20 30 40 50 60 70 80 90 100 
24 
22 
20 
18 
16 
14 
12 
10 
% of function evaluation 
log(fmin) 
CLPSO 
CMA−ES 
Figure : Results on function 2 CEC2015 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 16 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Experiments 
Test on CEC2015(shifted,rotated and non-seperable) 
0 10 20 30 40 50 60 70 80 90 100 
518 
516 
514 
512 
510 
508 
506 
504 
502 
500 
% of function evaluations 
fmin 
CMA−ES 
CLPSO 
Figure : Results on function 5 CEC2015 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 17 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Experiments 
Test on CEC2015(shifted,rotated and non-seperable) 
0 10 20 30 40 50 60 70 80 90 100 
612 
611 
610 
609 
608 
607 
606 
605 
604 
603 
% of function evaluation 
fmin 
CLPSO 
CMA−ES 
Figure : Results on function 6 CEC2015 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 18 / 19
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
.. 
. 
. .. 
. 
. .. 
. 
.. 
. 
.. 
. 
Conclusion 
Conclusion 
Applicable to problems in which many variables are correlated 
Good local search 
Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 19 / 19

More Related Content

Viewers also liked

Applications to Central Limit Theorem and Law of Large Numbers
Applications to Central Limit Theorem and Law of Large NumbersApplications to Central Limit Theorem and Law of Large Numbers
Applications to Central Limit Theorem and Law of Large NumbersUniversity of Salerno
 
Lecture: Monte Carlo Methods
Lecture: Monte Carlo MethodsLecture: Monte Carlo Methods
Lecture: Monte Carlo MethodsFrank Kienle
 
Monte Carlo Statistical Methods
Monte Carlo Statistical MethodsMonte Carlo Statistical Methods
Monte Carlo Statistical MethodsChristian Robert
 
Central limit theorem
Central limit theoremCentral limit theorem
Central limit theoremVijeesh Soman
 
Covariance Matrix Adaptation Evolution Strategy - CMA-ES
Covariance Matrix Adaptation Evolution Strategy - CMA-ESCovariance Matrix Adaptation Evolution Strategy - CMA-ES
Covariance Matrix Adaptation Evolution Strategy - CMA-ESOsama Salaheldin
 

Viewers also liked (6)

Applications to Central Limit Theorem and Law of Large Numbers
Applications to Central Limit Theorem and Law of Large NumbersApplications to Central Limit Theorem and Law of Large Numbers
Applications to Central Limit Theorem and Law of Large Numbers
 
Lecture: Monte Carlo Methods
Lecture: Monte Carlo MethodsLecture: Monte Carlo Methods
Lecture: Monte Carlo Methods
 
Monte Carlo Statistical Methods
Monte Carlo Statistical MethodsMonte Carlo Statistical Methods
Monte Carlo Statistical Methods
 
Central limit theorem
Central limit theoremCentral limit theorem
Central limit theorem
 
Monte carlo
Monte carloMonte carlo
Monte carlo
 
Covariance Matrix Adaptation Evolution Strategy - CMA-ES
Covariance Matrix Adaptation Evolution Strategy - CMA-ESCovariance Matrix Adaptation Evolution Strategy - CMA-ES
Covariance Matrix Adaptation Evolution Strategy - CMA-ES
 

Similar to Covariance Matrix Adaptation Evolution Strategy (CMA-ES)

Student manual
Student manualStudent manual
Student manualec931657
 
0580 s17 qp_42 exam - copy
0580 s17 qp_42 exam - copy0580 s17 qp_42 exam - copy
0580 s17 qp_42 exam - copymend Oyunchimeg
 
Presentation European Actuarial Journal conference 2016
Presentation European Actuarial Journal conference 2016Presentation European Actuarial Journal conference 2016
Presentation European Actuarial Journal conference 2016Thierry Moudiki
 
super-cheatsheet-artificial-intelligence.pdf
super-cheatsheet-artificial-intelligence.pdfsuper-cheatsheet-artificial-intelligence.pdf
super-cheatsheet-artificial-intelligence.pdfssuser089265
 
A Factor Graph Approach To Constrained Optimization
A Factor Graph Approach To Constrained OptimizationA Factor Graph Approach To Constrained Optimization
A Factor Graph Approach To Constrained OptimizationWendy Berg
 
Gnu octave help book 01 of 02
Gnu octave help book 01 of 02Gnu octave help book 01 of 02
Gnu octave help book 01 of 02Arun Umrao
 
Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...
Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...
Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...ssuserd6b1fd
 
Comparison GUM versus GUM+1
Comparison GUM  versus GUM+1Comparison GUM  versus GUM+1
Comparison GUM versus GUM+1Maurice Maeck
 
Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...
Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...
Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...Glenn Mallo
 
Hand book of Howard Anton calculus exercises 8th edition
Hand book of Howard Anton calculus exercises 8th editionHand book of Howard Anton calculus exercises 8th edition
Hand book of Howard Anton calculus exercises 8th editionPriSim
 
26 Machine Learning Unsupervised Fuzzy C-Means
26 Machine Learning Unsupervised Fuzzy C-Means26 Machine Learning Unsupervised Fuzzy C-Means
26 Machine Learning Unsupervised Fuzzy C-MeansAndres Mendez-Vazquez
 
Rapport d'analyse Dimensionality Reduction
Rapport d'analyse Dimensionality ReductionRapport d'analyse Dimensionality Reduction
Rapport d'analyse Dimensionality ReductionMatthieu Cisel
 

Similar to Covariance Matrix Adaptation Evolution Strategy (CMA-ES) (20)

Student manual
Student manualStudent manual
Student manual
 
0580 s17 qp_42 exam - copy
0580 s17 qp_42 exam - copy0580 s17 qp_42 exam - copy
0580 s17 qp_42 exam - copy
 
Presentation European Actuarial Journal conference 2016
Presentation European Actuarial Journal conference 2016Presentation European Actuarial Journal conference 2016
Presentation European Actuarial Journal conference 2016
 
super-cheatsheet-artificial-intelligence.pdf
super-cheatsheet-artificial-intelligence.pdfsuper-cheatsheet-artificial-intelligence.pdf
super-cheatsheet-artificial-intelligence.pdf
 
Imo 2008
Imo 2008 Imo 2008
Imo 2008
 
Imo2008 sl
Imo2008 slImo2008 sl
Imo2008 sl
 
COA_GHGInv_Full
COA_GHGInv_FullCOA_GHGInv_Full
COA_GHGInv_Full
 
A Factor Graph Approach To Constrained Optimization
A Factor Graph Approach To Constrained OptimizationA Factor Graph Approach To Constrained Optimization
A Factor Graph Approach To Constrained Optimization
 
Gnu octave help book 01 of 02
Gnu octave help book 01 of 02Gnu octave help book 01 of 02
Gnu octave help book 01 of 02
 
Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...
Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...
Notes for GNU Octave - Numerical Programming - for Students 01 of 02 by Arun ...
 
Comparison GUM versus GUM+1
Comparison GUM  versus GUM+1Comparison GUM  versus GUM+1
Comparison GUM versus GUM+1
 
Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...
Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...
Rely on Everyday Words & other good stuff no more (Chapter 3 of superseded SE...
 
Hand book of Howard Anton calculus exercises 8th edition
Hand book of Howard Anton calculus exercises 8th editionHand book of Howard Anton calculus exercises 8th edition
Hand book of Howard Anton calculus exercises 8th edition
 
Kk2518251830
Kk2518251830Kk2518251830
Kk2518251830
 
Kk2518251830
Kk2518251830Kk2518251830
Kk2518251830
 
Report_Jeremy_Berard
Report_Jeremy_BerardReport_Jeremy_Berard
Report_Jeremy_Berard
 
26 Machine Learning Unsupervised Fuzzy C-Means
26 Machine Learning Unsupervised Fuzzy C-Means26 Machine Learning Unsupervised Fuzzy C-Means
26 Machine Learning Unsupervised Fuzzy C-Means
 
9702 w03 ms_all
9702 w03 ms_all9702 w03 ms_all
9702 w03 ms_all
 
Support Vector Machines
Support Vector MachinesSupport Vector Machines
Support Vector Machines
 
Rapport d'analyse Dimensionality Reduction
Rapport d'analyse Dimensionality ReductionRapport d'analyse Dimensionality Reduction
Rapport d'analyse Dimensionality Reduction
 

Recently uploaded

CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfCCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfAsst.prof M.Gokilavani
 
Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.eptoze12
 
Arduino_CSE ece ppt for working and principal of arduino.ppt
Arduino_CSE ece ppt for working and principal of arduino.pptArduino_CSE ece ppt for working and principal of arduino.ppt
Arduino_CSE ece ppt for working and principal of arduino.pptSAURABHKUMAR892774
 
Risk Assessment For Installation of Drainage Pipes.pdf
Risk Assessment For Installation of Drainage Pipes.pdfRisk Assessment For Installation of Drainage Pipes.pdf
Risk Assessment For Installation of Drainage Pipes.pdfROCENODodongVILLACER
 
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidmain PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidNikhilNagaraju
 
Application of Residue Theorem to evaluate real integrations.pptx
Application of Residue Theorem to evaluate real integrations.pptxApplication of Residue Theorem to evaluate real integrations.pptx
Application of Residue Theorem to evaluate real integrations.pptx959SahilShah
 
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube ExchangerStudy on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube ExchangerAnamika Sarkar
 
8251 universal synchronous asynchronous receiver transmitter
8251 universal synchronous asynchronous receiver transmitter8251 universal synchronous asynchronous receiver transmitter
8251 universal synchronous asynchronous receiver transmitterShivangiSharma879191
 
Churning of Butter, Factors affecting .
Churning of Butter, Factors affecting  .Churning of Butter, Factors affecting  .
Churning of Butter, Factors affecting .Satyam Kumar
 
Correctly Loading Incremental Data at Scale
Correctly Loading Incremental Data at ScaleCorrectly Loading Incremental Data at Scale
Correctly Loading Incremental Data at ScaleAlluxio, Inc.
 
Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...VICTOR MAESTRE RAMIREZ
 
Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024hassan khalil
 
Introduction to Machine Learning Unit-3 for II MECH
Introduction to Machine Learning Unit-3 for II MECHIntroduction to Machine Learning Unit-3 for II MECH
Introduction to Machine Learning Unit-3 for II MECHC Sai Kiran
 
Sachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective IntroductionSachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective IntroductionDr.Costas Sachpazis
 
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxDecoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxJoão Esperancinha
 
Call Girls Narol 7397865700 Independent Call Girls
Call Girls Narol 7397865700 Independent Call GirlsCall Girls Narol 7397865700 Independent Call Girls
Call Girls Narol 7397865700 Independent Call Girlsssuser7cb4ff
 
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort serviceGurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort servicejennyeacort
 
Heart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptxHeart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptxPoojaBan
 

Recently uploaded (20)

CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdfCCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
CCS355 Neural Network & Deep Learning Unit II Notes with Question bank .pdf
 
Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.Oxy acetylene welding presentation note.
Oxy acetylene welding presentation note.
 
Arduino_CSE ece ppt for working and principal of arduino.ppt
Arduino_CSE ece ppt for working and principal of arduino.pptArduino_CSE ece ppt for working and principal of arduino.ppt
Arduino_CSE ece ppt for working and principal of arduino.ppt
 
Risk Assessment For Installation of Drainage Pipes.pdf
Risk Assessment For Installation of Drainage Pipes.pdfRisk Assessment For Installation of Drainage Pipes.pdf
Risk Assessment For Installation of Drainage Pipes.pdf
 
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidmain PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfid
 
Application of Residue Theorem to evaluate real integrations.pptx
Application of Residue Theorem to evaluate real integrations.pptxApplication of Residue Theorem to evaluate real integrations.pptx
Application of Residue Theorem to evaluate real integrations.pptx
 
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube ExchangerStudy on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
Study on Air-Water & Water-Water Heat Exchange in a Finned Tube Exchanger
 
8251 universal synchronous asynchronous receiver transmitter
8251 universal synchronous asynchronous receiver transmitter8251 universal synchronous asynchronous receiver transmitter
8251 universal synchronous asynchronous receiver transmitter
 
Churning of Butter, Factors affecting .
Churning of Butter, Factors affecting  .Churning of Butter, Factors affecting  .
Churning of Butter, Factors affecting .
 
Correctly Loading Incremental Data at Scale
Correctly Loading Incremental Data at ScaleCorrectly Loading Incremental Data at Scale
Correctly Loading Incremental Data at Scale
 
Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...Software and Systems Engineering Standards: Verification and Validation of Sy...
Software and Systems Engineering Standards: Verification and Validation of Sy...
 
POWER SYSTEMS-1 Complete notes examples
POWER SYSTEMS-1 Complete notes  examplesPOWER SYSTEMS-1 Complete notes  examples
POWER SYSTEMS-1 Complete notes examples
 
Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024Architect Hassan Khalil Portfolio for 2024
Architect Hassan Khalil Portfolio for 2024
 
Introduction to Machine Learning Unit-3 for II MECH
Introduction to Machine Learning Unit-3 for II MECHIntroduction to Machine Learning Unit-3 for II MECH
Introduction to Machine Learning Unit-3 for II MECH
 
Design and analysis of solar grass cutter.pdf
Design and analysis of solar grass cutter.pdfDesign and analysis of solar grass cutter.pdf
Design and analysis of solar grass cutter.pdf
 
Sachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective IntroductionSachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
Sachpazis Costas: Geotechnical Engineering: A student's Perspective Introduction
 
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxDecoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
 
Call Girls Narol 7397865700 Independent Call Girls
Call Girls Narol 7397865700 Independent Call GirlsCall Girls Narol 7397865700 Independent Call Girls
Call Girls Narol 7397865700 Independent Call Girls
 
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort serviceGurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
Gurgaon ✡️9711147426✨Call In girls Gurgaon Sector 51 escort service
 
Heart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptxHeart Disease Prediction using machine learning.pptx
Heart Disease Prediction using machine learning.pptx
 

Covariance Matrix Adaptation Evolution Strategy (CMA-ES)

  • 1. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . . . Covariance Matrix Adaptation Evolution Strategies(CMA-ES) Hossein Abedi Evolutionary Computation Autumn 2014 Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 1 / 19
  • 2. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Overview 1. Introduction 2. Selection and Recombination 3. Adaptation of covariance matrix 4. Step size control 5. Experiments 6. Conclusion Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 2 / 19
  • 3. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Introduction Idea Introduced by Hansen and Ostermeier in 2001 The idea: Figure : Movement toward a minimum through 3 generations Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 3 / 19
  • 4. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Selection and Recombination Generating the children New points are sampled normally distributed: Xi M + Ni (0,C ), for i=1,..., Figure : Different shapes of C as a hyperelipsoid in 2D Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 4 / 19
  • 5. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Selection and Recombination Selection and Recombination The mean vector M 2 ℜn is calculated Σas the weighted average of the best candidate solutions: M= i=1 wiXi : Where: Σ i=1 wi = 1 w1 ⩾ w2 ⩾ ::: ⩾ w 0 f (X1:) ⩽ f (X2:) ⩽ ::: ⩽ f (X:) eff = ( jjwjj1 jjwjj2 )2 = Σ 1 i=1 w2 i Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 5 / 19
  • 6. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Estimating the covariance matrix from scratch For the sake of simplicity set (g) = 1 Estimating distribution within the population: C(g+1) emp = 1 1 Σ i=1(X(g+1) i 1 Σ j=1 Xj )(X(g+1) i 1 Σ j=1 Xj )T Estimating distribution of sampled steps: C(g+1) = 1 Σ i=1(X(g+1) i M(g))(X(g+1) i M(g))T Where: The sampled steps are: X(g+1) M(g) i Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 6 / 19
  • 7. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Estimating the covariance matrix Estimating distribution of the most successful steps: C(g+1) = 1 Σ i=1 wi (X(g+1) i : M(g))(X(g+1) i : M(g))T Estimation of Multivariate Normal Algorithm(ENMA): C(g+1) = 1 Σ i=1(X(g+1) i : M(g+1) enma )(X(g+1) i : M(g+1) enma )T Where: M(g+1) enma = 1 Σ j=1 X(g+1) j : Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 7 / 19
  • 8. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Estimating the covariance matrix Comparison: Figure : Covariance matrix estimation on f (x1; x2) = Σ2 i=1 xi Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 8 / 19
  • 9. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Rank update Smaller means faster but less global search To give recent generations a higher weight, consider a leraning rate c and the equation below: C(g+1) = (1 c)C(g) + c 1 C(g+1) ((g))2 Where: 1 c is called the time back horizon Figure : Example of exponential smoothing Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 9 / 19
  • 10. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Rank update C(g+1) = (1 c)C(g) + c 1 Σ i=1 wiOP(X(g+1) i : M(g) (g) ) Where: OP(y) = yyT Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 10 / 19
  • 11. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Rank one update Evolution Path (Pc 2 ℜn): sum of consecutive steps: M(g+1)M(g) (g) + M(g)M(g1) (g1) + ::: Figure : Evolution path N(0; I )y1 + N(0; I )y2 + ::: + N(0; I )yg N(0; Σg i=1 yi yT i ) Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 11 / 19
  • 12. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Rank one update Using exponential smoothing: P(g+1) c = (1 cc )P(g) c + √ cc(2 cc )eff M(g+1)M(g) (g) Wher√e: cc(2 cc )eff is a scaling factor such that :P(g+1) c N(0; C) So rank one update with sign is : C(g+1) = (1 c1)C(g) + c1OP(P(g+1) c ) Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 12 / 19
  • 13. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Adaptation of covariance matrix Cumulation C(g+1) = (1 c1 c)C(g) + c1(y(g+1) c )(P(g+1) c )T + :::c 1 Σ i=1 wiOP(X(g+1) i : M(g) (g) ) Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 13 / 19
  • 14. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Step size control Step size control Using the evolution path for adapting the stepsize Figure : Different evolution path senarios for 6 consecutive mean vectors (g+1) = (g) exp ( c d ( jjp(g+1) jj EjjN(0;I )jj 1)) Where: p(g+1) = (1 c)p(g) + √ c(2 c)eff (C(g))1 2 M(g+1)M(g) (g) Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 14 / 19
  • 15. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Experiments Test on seperable and non rotated 0 200 400 600 800 1000 1200 1400 1600 1800 2000 14 12 10 8 6 4 2 0 0.01*function evauations fmin CLPSO CMA−ES Figure : Results on Ackley test function Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 15 / 19
  • 16. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Experiments Test on CEC2015(shifted,rotated and non-seperable) 0 10 20 30 40 50 60 70 80 90 100 24 22 20 18 16 14 12 10 % of function evaluation log(fmin) CLPSO CMA−ES Figure : Results on function 2 CEC2015 Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 16 / 19
  • 17. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Experiments Test on CEC2015(shifted,rotated and non-seperable) 0 10 20 30 40 50 60 70 80 90 100 518 516 514 512 510 508 506 504 502 500 % of function evaluations fmin CMA−ES CLPSO Figure : Results on function 5 CEC2015 Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 17 / 19
  • 18. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Experiments Test on CEC2015(shifted,rotated and non-seperable) 0 10 20 30 40 50 60 70 80 90 100 612 611 610 609 608 607 606 605 604 603 % of function evaluation fmin CLPSO CMA−ES Figure : Results on function 6 CEC2015 Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 18 / 19
  • 19. .. . . .. . . .. . .. . .. . .. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .. . .. . .. . Conclusion Conclusion Applicable to problems in which many variables are correlated Good local search Hossein Abedi (Evolutionary Computation) CMA-ES Autumn 2014 19 / 19