SlideShare una empresa de Scribd logo
1 de 25
Descargar para leer sin conexión
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
1 of 25
On Partitioned Fitness-Distributions
of Genetic Operators for Predicting
GA Performance
WSC6
6th Online World Conference on Soft Computing in
Industrial Applications
Departamento de Lenguajes y
Ciencias de la Computación
Universidad de Málaga
Rafael Nogueras, Carlos Cotta
September 10-24, 2001
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
2 of 25
Outline of the
Presentation
1. Introduction
2. Virtual Genetic Algorithms
3. Enhancing the Model
4. Experimental Results
5. Conclusions
Outline of the Presentation
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
3 of 25
• The computational cost of running a GA can be
overwhelming.
• Maximizing GA effectiveness  minimizing the
number of fitness calculations needed to achieve
a certain solution.
• A careful instantiation of the basic template is
required (popsize, pX, pm, …).
• Tools for predicting GA performance are
necessary:
▪ Information-based
▪ Fitness-based
Introduction
Introduction
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
4 of 25
Virtual Genetic Algorithms
• Virtual Genetic Algorithms (VGAs) constitute a
statistical approach aimed at GA-performance
prediction.
• VGAs vs GAs: individuals do not carry
genetic information; they are used as place-
holders for fitness information.
• Fitness is transmitted from parents to
offspring, using a statistical model to simulate
the effects of genetic operators.
• This statistical model is utilized to predict the
fitness of a new individual in terms of the
fitnesses of its parents.
Virtual Genetic
Algorithms
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
5 of 25
Sketch of a VGA
Initialize population
Termination
condition?
Select parents
Apply virtual
operators
Insert in population
NO
YES
Statistical model
fo = fo(fp,P)
Parameters of the model
Inferred off-line
Virtual Genetic
Algorithms
Sketch of a VGA
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
6 of 25
• Based on a linear regression approach for
each genetic operator.
• Input: a dataset comprising N pairs (fp, fo).
– fp = (mean) fitness of random (pair of) solution(s).
– fo = fitness of the solution obtained by applying an
genetic operator to the solution(s) above.
• Output: a tuple (a, b, sa, sb).
– a, b = Least-squares fit of fo = a+ bfp to the dataset.
– sa, sb = Least-squares fit of so = sa + sb fp, where
so and fp are the standard deviation of fo and
the average of fp when grouped in bins of K
pairs.
The Basic VGA
Virtual Genetic
Algorithms
The Basic VGA
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
7 of 25
The Basic VGA
• The VGA predicts the descendant fitness using
a Gaussian distribution: fo=N(a+bfp, sa+ sbfp) .
• This approach just provides discrete results.
OneMax Brachystochrone
• Two major reasons: acquisition of the dataset
and asymmetry of the distribution.
Virtual Genetic
Algorithms
The Basic VGA
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
8 of 25
Problems of the Basic VGA
Acquisition of the dataset:
Random sampling provides significant
information in simple functions, but not in
general.
OneMax Brachystochrone
Virtual Genetic
Algorithms
Problems of the Basic
VGA
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
9 of 25
Problems of the Basic VGA
Asymmetry of the distribution:
The Gaussian distribution is symmetric, but
the fitness distribution can be asymmetric.
Brachystochrone Rastrigin Function
Virtual Genetic
Algorithms
Problems of the Basic
VGA
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
10 of 25
The L-VGA1
• An enhanced version of the VGA was obtained.
• Acquisition of the dataset: use a dataset
obtained by means of a pilot run of a GA.
– Computational cost similar to random sampling.
– The dataset covers a larger region of the fitness space.
• Asymmetry: find the parameters of the linear fit
fo=a+bfp, and split the dataset in two datasets:
– U = { (fp,fo) | fo > a + bfp }
– L = { (fp,fo) | fo  a + bfp }
Virtual Genetic
Algorithms
Enhancing the Model
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
11 of 25
The L-VGA1
• Input: a dataset comprising N pairs (fp, fo).
• Output: a tuple
• COMPUTE-MODEL:
– a, b = Least-squares fit of fo = a+ bfp to the dataset.
– = Least-squares fit of so = sa + sb fp, where so
and fp are the standard deviation of fo and the average of fp
when grouped in bins of K pairs For U or L.
– = Least-squares fit of mo = la + lb fp, where mo
and fp are the minimum of fo and the average of fp when
grouped in bins of K pairs for U or L.
– p = relative weight of U vs. L.
)
,
,
,
,
,
,
,
, p
b
a L
b
L
a
U
b
U
a
L
b
L
a
U
b
U
a l
l
l
l
s
s
s
s


(
)
, |
| L
U
b
L
U
a s
(s
)
, |
| L
U
b
L
U
a l
(l
Virtual Genetic
Algorithms
Enhancing the Model
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
12 of 25
Some Results
Brachystochrone
Some Results
Virtual Genetic
Algorithms
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
13 of 25
Some Results
Rastrigin Function
Some Results
Virtual Genetic
Algorithms
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
14 of 25
Some Results
Flowshop Scheduling
Some Results
Virtual Genetic
Algorithms
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
15 of 25
Partitioning the Dataset
Problem of the L-VGA1:
• The underlying linear model is somewhat
inflexible.
• Different behavior of the GA in different
regions of fitness space.
• The chances for obtaining a descendant better
than the parent are higher when the latter has a
lower quality, for example, during some initial
stages of the GA run.
Partitioning the
Dataset
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
16 of 25
Partitioning the Dataset
Solutions:
• Use a higher-degree polynomial model, but it
does not offer a conclusive improvement for a
low-degree polynomial.
• Partition of the dataset into several prediction
intervals, and extract a model for each of them.
Flowshop Scheduling
Brachystochrone
Partitioning the
Dataset
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
17 of 25
The L-VGAM
Input: a dataset comprising N pairs (fp, fo).
Output: M tuples
• Sort the dataset according to fp, and partition it into M
subsets Si.
• Si = [Li, Ri), Li = Ri-1, L1 = min(fp), RM = max(fp)+1, i.e.,
Si = { ( fp,fo ) | Li  fp < Ri }
• Mi = COMPUTE MODEL for Si, i=1..M, where:
Mi =
• OUTPUT the complete model M = {M1,…, MM}
)
,
,
,
,
,
,
,
, p
b
a L
b
L
a
U
b
U
a
L
b
L
a
U
b
U
a l
l
l
l
s
s
s
s


(
)
,
,
,
,
,
,
,
, p
b
a L
b
L
a
U
b
U
a
L
b
L
a
U
b
U
a l
l
l
l
s
s
s
s


(
The Enhanced VGA
Partitioning the
Dataset
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
18 of 25
Experimental Results
• Experiments have been done with an elitist
generational GA/VGA (popsize = 100, pX = .9,
pm=.01, maxevals=10000) and PGA/PVGA
(popsize = 300/400, pX = .9, pm = .01, maxevals =
= 20000) using ranking selection.
• All results correspond to series of 50 runs.
• Three problems have been considered: the Rastrigin
function, the Fletcher-Powell function, and flowshop
scheduling.
• Issues tackled:
– Goodness of quantitative estimation
– Goodness of qualitative estimation
– Computational speed-up
Experimental
Results
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
19 of 25
Quantitative Estimations
Fletcher-Powell Function
Experimental
Results
Quantitative Estimations
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
20 of 25
Quantitative Estimations
Flowshop Scheduling
Experimental
Results
Quantitative Estimations
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
21 of 25
Qualitative Estimations
Three different GA configurations are compared
on the Rastrigin Function:
A = {popsize=100, pX=.9, pm=.01}
B = {popsize=50, pX=.5, pm=.10}
C = {popsize=80, pX=.7, pm=.05}
Real GA Virtual GA
Relative properties are preserved.
Experimental
Results
Qualitative Estimations
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
22 of 25
Qualitative Estimations
Four different Parallel GA (PGA) configurations are
compared on a flowshop scheduling problem:
A = {GAs=1}, C = {GAs=4} B = {GAs=2}, D = {GAs=8}
Real PGA Virtual PGA
Relative properties are preserved.
Experimental
Results
Qualitative Estimations
(ring topology)
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
23 of 25
Qualitative Estimations
Three different PGA configurations are compared
on a flowshop scheduling problem:
A = {GAs = 2, migfreq = 1000, migrate-best}
B = {GAs = 2, migfreq = 300, migrate-random}
C = {GAs = 4, migfreq = 500, migrate-random}
Real PGA Virtual PGA
Relative properties are preserved.
Experimental
Results
Qualitative Estimations
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
24 of 25
Speed-Up
The larger the problem instance, the higher the
speed-up of the VGA vs. GA.
Experimental
Results
Speed-Up
On Partitioned Fitness-
Distributions of Genetic
Operators for Predicting
GA Performance
Outline of the
Presentation
Introduction
Virtual Genetic
Algorithms
Experimental
Results
Conclusions
Sketch of a VGA
The Basic VGA
Problems of the Basic
VGA
Enhancing the Model
Quantitative Estimations
Qualitative Estimations
Speed-Up
Some Results
Partitioning the
Dataset
The Enhanced VGA
9/26/2021
25 of 25
Conclusions
• VGAs can be a useful tool for predicting GA
performance.
• An enhanced VGA model has been proposed
to grasp some properties of the fitness
distribution such as asymmetry, statistical
anisotropy, or the different behaviour of GA.
• The obtained results show:
✓ a much more accurate performance estimation
✓ high computational speed-ups
• Future work: study other schemes for
partitioning the dataset and other non-linear
statistical models
Conclusions

Más contenido relacionado

Similar a Partitioned Fitness Distributions Predict GA Performance

Bayesian Optimization for Balancing Metrics in Recommender Systems
Bayesian Optimization for Balancing Metrics in Recommender SystemsBayesian Optimization for Balancing Metrics in Recommender Systems
Bayesian Optimization for Balancing Metrics in Recommender SystemsViral Gupta
 
Classifiers Optimization Using Swarm Algorithms
Classifiers Optimization Using Swarm AlgorithmsClassifiers Optimization Using Swarm Algorithms
Classifiers Optimization Using Swarm AlgorithmsAboul Ella Hassanien
 
Data Analytics and Distribution with Power BI
 Data Analytics and Distribution with Power BI Data Analytics and Distribution with Power BI
Data Analytics and Distribution with Power BIdesertislesql
 
Using Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning ModelsUsing Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning ModelsScott Clark
 
Using Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning ModelsUsing Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning ModelsSigOpt
 
Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...
Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...
Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...Varun Ojha
 
統計在半導體產業的應用 -- Basic Statistic Methods
統計在半導體產業的應用 -- Basic Statistic Methods統計在半導體產業的應用 -- Basic Statistic Methods
統計在半導體產業的應用 -- Basic Statistic Methodschunhung chou
 
Measure phase lean six sigma tollgate template
Measure phase   lean six sigma tollgate templateMeasure phase   lean six sigma tollgate template
Measure phase lean six sigma tollgate templateSteven Bonacorsi
 
Measure phase lean six sigma tollgate template
Measure phase   lean six sigma tollgate templateMeasure phase   lean six sigma tollgate template
Measure phase lean six sigma tollgate templateSteven Bonacorsi
 
SAP HANA SPS09 - Predictive Analysis Library
SAP HANA SPS09 - Predictive Analysis LibrarySAP HANA SPS09 - Predictive Analysis Library
SAP HANA SPS09 - Predictive Analysis LibrarySAP Technology
 
EPC2152 half bridge Monolithic GaN IC
EPC2152 half bridge Monolithic GaN ICEPC2152 half bridge Monolithic GaN IC
EPC2152 half bridge Monolithic GaN ICsystem_plus
 
7. space the estimation aid for bringing agile delivery predictability - p...
7. space   the estimation aid for bringing agile delivery predictability  - p...7. space   the estimation aid for bringing agile delivery predictability  - p...
7. space the estimation aid for bringing agile delivery predictability - p...Nesma
 
ams’ NanEye Mini Camera
ams’ NanEye Mini Cameraams’ NanEye Mini Camera
ams’ NanEye Mini Camerasystem_plus
 
IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...
IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...
IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...IRJET Journal
 
Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...
Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...
Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...IRJET Journal
 
Enhancing growth curve approach using cgpann for predicting
Enhancing growth curve approach using cgpann for predictingEnhancing growth curve approach using cgpann for predicting
Enhancing growth curve approach using cgpann for predictingJawad Khan
 
A Pragmatic Approach to Semantic Repositories Benchmarking
A Pragmatic Approach to Semantic Repositories BenchmarkingA Pragmatic Approach to Semantic Repositories Benchmarking
A Pragmatic Approach to Semantic Repositories BenchmarkingDhaval Thakker
 

Similar a Partitioned Fitness Distributions Predict GA Performance (20)

Bayesian Optimization for Balancing Metrics in Recommender Systems
Bayesian Optimization for Balancing Metrics in Recommender SystemsBayesian Optimization for Balancing Metrics in Recommender Systems
Bayesian Optimization for Balancing Metrics in Recommender Systems
 
Classifiers Optimization Using Swarm Algorithms
Classifiers Optimization Using Swarm AlgorithmsClassifiers Optimization Using Swarm Algorithms
Classifiers Optimization Using Swarm Algorithms
 
Ssbse12b.ppt
Ssbse12b.pptSsbse12b.ppt
Ssbse12b.ppt
 
Data Analytics and Distribution with Power BI
 Data Analytics and Distribution with Power BI Data Analytics and Distribution with Power BI
Data Analytics and Distribution with Power BI
 
Using Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning ModelsUsing Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning Models
 
Using Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning ModelsUsing Bayesian Optimization to Tune Machine Learning Models
Using Bayesian Optimization to Tune Machine Learning Models
 
Ijcai 2020
Ijcai 2020Ijcai 2020
Ijcai 2020
 
Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...
Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...
Dimensionality Reduction and Prediction of the Protein Macromolecule Dissolut...
 
統計在半導體產業的應用 -- Basic Statistic Methods
統計在半導體產業的應用 -- Basic Statistic Methods統計在半導體產業的應用 -- Basic Statistic Methods
統計在半導體產業的應用 -- Basic Statistic Methods
 
Measure phase lean six sigma tollgate template
Measure phase   lean six sigma tollgate templateMeasure phase   lean six sigma tollgate template
Measure phase lean six sigma tollgate template
 
Measure phase lean six sigma tollgate template
Measure phase   lean six sigma tollgate templateMeasure phase   lean six sigma tollgate template
Measure phase lean six sigma tollgate template
 
SAP HANA SPS09 - Predictive Analysis Library
SAP HANA SPS09 - Predictive Analysis LibrarySAP HANA SPS09 - Predictive Analysis Library
SAP HANA SPS09 - Predictive Analysis Library
 
Tree building 2
Tree building 2Tree building 2
Tree building 2
 
EPC2152 half bridge Monolithic GaN IC
EPC2152 half bridge Monolithic GaN ICEPC2152 half bridge Monolithic GaN IC
EPC2152 half bridge Monolithic GaN IC
 
7. space the estimation aid for bringing agile delivery predictability - p...
7. space   the estimation aid for bringing agile delivery predictability  - p...7. space   the estimation aid for bringing agile delivery predictability  - p...
7. space the estimation aid for bringing agile delivery predictability - p...
 
ams’ NanEye Mini Camera
ams’ NanEye Mini Cameraams’ NanEye Mini Camera
ams’ NanEye Mini Camera
 
IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...
IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...
IRJET- A Review- FPGA based Architectures for Image Capturing Consequently Pr...
 
Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...
Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...
Operation Sequencing and Machining Parameter Selection in CAPP for Cylindrica...
 
Enhancing growth curve approach using cgpann for predicting
Enhancing growth curve approach using cgpann for predictingEnhancing growth curve approach using cgpann for predicting
Enhancing growth curve approach using cgpann for predicting
 
A Pragmatic Approach to Semantic Repositories Benchmarking
A Pragmatic Approach to Semantic Repositories BenchmarkingA Pragmatic Approach to Semantic Repositories Benchmarking
A Pragmatic Approach to Semantic Repositories Benchmarking
 

Más de 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
 
Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...
Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...
Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...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
 
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
 
Analyzing Meme Propagation in Multimemetic Algorithms
Analyzing Meme Propagation in Multimemetic AlgorithmsAnalyzing Meme Propagation in Multimemetic Algorithms
Analyzing Meme Propagation in 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
 
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
 

Más de Rafael Nogueras (11)

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...
 
Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...
Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...
Sensitivity Analysis of Checkpointing Strategies for Multimemetic Algorithms ...
 
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
 
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...
 
Analyzing Meme Propagation in Multimemetic Algorithms
Analyzing Meme Propagation in Multimemetic AlgorithmsAnalyzing Meme Propagation in Multimemetic Algorithms
Analyzing Meme Propagation in 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 ...
 
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
 

Último

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
 
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
 
Environmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial BiosensorEnvironmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial Biosensorsonawaneprad
 
Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024AyushiRastogi48
 
Four Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptFour Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptJoemSTuliba
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxpriyankatabhane
 
Pests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdfPests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdfPirithiRaju
 
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
 
Pests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdfPests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdfPirithiRaju
 
User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)Columbia Weather Systems
 
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptxTHE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptxNandakishor Bhaurao Deshmukh
 
Transposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.pptTransposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.pptArshadWarsi13
 
Call Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 Genuine
Call Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 GenuineCall Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 Genuine
Call Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 Genuinethapagita
 
Good agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptxGood agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptxSimeonChristian
 
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...Universidade Federal de Sergipe - UFS
 
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
 
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...lizamodels9
 
Pests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdfPests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdfPirithiRaju
 
Neurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 trNeurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 trssuser06f238
 
User Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationUser Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationColumbia Weather Systems
 

Último (20)

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🔝
 
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
 
Environmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial BiosensorEnvironmental Biotechnology Topic:- Microbial Biosensor
Environmental Biotechnology Topic:- Microbial Biosensor
 
Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024Vision and reflection on Mining Software Repositories research in 2024
Vision and reflection on Mining Software Repositories research in 2024
 
Four Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.pptFour Spheres of the Earth Presentation.ppt
Four Spheres of the Earth Presentation.ppt
 
Speech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptxSpeech, hearing, noise, intelligibility.pptx
Speech, hearing, noise, intelligibility.pptx
 
Pests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdfPests of Bengal gram_Identification_Dr.UPR.pdf
Pests of Bengal gram_Identification_Dr.UPR.pdf
 
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
 
Pests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdfPests of soyabean_Binomics_IdentificationDr.UPR.pdf
Pests of soyabean_Binomics_IdentificationDr.UPR.pdf
 
User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)User Guide: Orion™ Weather Station (Columbia Weather Systems)
User Guide: Orion™ Weather Station (Columbia Weather Systems)
 
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptxTHE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
THE ROLE OF PHARMACOGNOSY IN TRADITIONAL AND MODERN SYSTEM OF MEDICINE.pptx
 
Transposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.pptTransposable elements in prokaryotes.ppt
Transposable elements in prokaryotes.ppt
 
Call Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 Genuine
Call Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 GenuineCall Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 Genuine
Call Girls in Majnu Ka Tilla Delhi 🔝9711014705🔝 Genuine
 
Good agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptxGood agricultural practices 3rd year bpharm. herbal drug technology .pptx
Good agricultural practices 3rd year bpharm. herbal drug technology .pptx
 
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
REVISTA DE BIOLOGIA E CIÊNCIAS DA TERRA ISSN 1519-5228 - Artigo_Bioterra_V24_...
 
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
 
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
Best Call Girls In Sector 29 Gurgaon❤️8860477959 EscorTs Service In 24/7 Delh...
 
Pests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdfPests of jatropha_Bionomics_identification_Dr.UPR.pdf
Pests of jatropha_Bionomics_identification_Dr.UPR.pdf
 
Neurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 trNeurodevelopmental disorders according to the dsm 5 tr
Neurodevelopmental disorders according to the dsm 5 tr
 
User Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather StationUser Guide: Capricorn FLX™ Weather Station
User Guide: Capricorn FLX™ Weather Station
 

Partitioned Fitness Distributions Predict GA Performance

  • 1. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 1 of 25 On Partitioned Fitness-Distributions of Genetic Operators for Predicting GA Performance WSC6 6th Online World Conference on Soft Computing in Industrial Applications Departamento de Lenguajes y Ciencias de la Computación Universidad de Málaga Rafael Nogueras, Carlos Cotta September 10-24, 2001
  • 2. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 2 of 25 Outline of the Presentation 1. Introduction 2. Virtual Genetic Algorithms 3. Enhancing the Model 4. Experimental Results 5. Conclusions Outline of the Presentation
  • 3. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 3 of 25 • The computational cost of running a GA can be overwhelming. • Maximizing GA effectiveness  minimizing the number of fitness calculations needed to achieve a certain solution. • A careful instantiation of the basic template is required (popsize, pX, pm, …). • Tools for predicting GA performance are necessary: ▪ Information-based ▪ Fitness-based Introduction Introduction
  • 4. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 4 of 25 Virtual Genetic Algorithms • Virtual Genetic Algorithms (VGAs) constitute a statistical approach aimed at GA-performance prediction. • VGAs vs GAs: individuals do not carry genetic information; they are used as place- holders for fitness information. • Fitness is transmitted from parents to offspring, using a statistical model to simulate the effects of genetic operators. • This statistical model is utilized to predict the fitness of a new individual in terms of the fitnesses of its parents. Virtual Genetic Algorithms
  • 5. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 5 of 25 Sketch of a VGA Initialize population Termination condition? Select parents Apply virtual operators Insert in population NO YES Statistical model fo = fo(fp,P) Parameters of the model Inferred off-line Virtual Genetic Algorithms Sketch of a VGA
  • 6. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 6 of 25 • Based on a linear regression approach for each genetic operator. • Input: a dataset comprising N pairs (fp, fo). – fp = (mean) fitness of random (pair of) solution(s). – fo = fitness of the solution obtained by applying an genetic operator to the solution(s) above. • Output: a tuple (a, b, sa, sb). – a, b = Least-squares fit of fo = a+ bfp to the dataset. – sa, sb = Least-squares fit of so = sa + sb fp, where so and fp are the standard deviation of fo and the average of fp when grouped in bins of K pairs. The Basic VGA Virtual Genetic Algorithms The Basic VGA
  • 7. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 7 of 25 The Basic VGA • The VGA predicts the descendant fitness using a Gaussian distribution: fo=N(a+bfp, sa+ sbfp) . • This approach just provides discrete results. OneMax Brachystochrone • Two major reasons: acquisition of the dataset and asymmetry of the distribution. Virtual Genetic Algorithms The Basic VGA
  • 8. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 8 of 25 Problems of the Basic VGA Acquisition of the dataset: Random sampling provides significant information in simple functions, but not in general. OneMax Brachystochrone Virtual Genetic Algorithms Problems of the Basic VGA
  • 9. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 9 of 25 Problems of the Basic VGA Asymmetry of the distribution: The Gaussian distribution is symmetric, but the fitness distribution can be asymmetric. Brachystochrone Rastrigin Function Virtual Genetic Algorithms Problems of the Basic VGA
  • 10. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 10 of 25 The L-VGA1 • An enhanced version of the VGA was obtained. • Acquisition of the dataset: use a dataset obtained by means of a pilot run of a GA. – Computational cost similar to random sampling. – The dataset covers a larger region of the fitness space. • Asymmetry: find the parameters of the linear fit fo=a+bfp, and split the dataset in two datasets: – U = { (fp,fo) | fo > a + bfp } – L = { (fp,fo) | fo  a + bfp } Virtual Genetic Algorithms Enhancing the Model
  • 11. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 11 of 25 The L-VGA1 • Input: a dataset comprising N pairs (fp, fo). • Output: a tuple • COMPUTE-MODEL: – a, b = Least-squares fit of fo = a+ bfp to the dataset. – = Least-squares fit of so = sa + sb fp, where so and fp are the standard deviation of fo and the average of fp when grouped in bins of K pairs For U or L. – = Least-squares fit of mo = la + lb fp, where mo and fp are the minimum of fo and the average of fp when grouped in bins of K pairs for U or L. – p = relative weight of U vs. L. ) , , , , , , , , p b a L b L a U b U a L b L a U b U a l l l l s s s s   ( ) , | | L U b L U a s (s ) , | | L U b L U a l (l Virtual Genetic Algorithms Enhancing the Model
  • 12. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 12 of 25 Some Results Brachystochrone Some Results Virtual Genetic Algorithms
  • 13. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 13 of 25 Some Results Rastrigin Function Some Results Virtual Genetic Algorithms
  • 14. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 14 of 25 Some Results Flowshop Scheduling Some Results Virtual Genetic Algorithms
  • 15. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 15 of 25 Partitioning the Dataset Problem of the L-VGA1: • The underlying linear model is somewhat inflexible. • Different behavior of the GA in different regions of fitness space. • The chances for obtaining a descendant better than the parent are higher when the latter has a lower quality, for example, during some initial stages of the GA run. Partitioning the Dataset
  • 16. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 16 of 25 Partitioning the Dataset Solutions: • Use a higher-degree polynomial model, but it does not offer a conclusive improvement for a low-degree polynomial. • Partition of the dataset into several prediction intervals, and extract a model for each of them. Flowshop Scheduling Brachystochrone Partitioning the Dataset
  • 17. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 17 of 25 The L-VGAM Input: a dataset comprising N pairs (fp, fo). Output: M tuples • Sort the dataset according to fp, and partition it into M subsets Si. • Si = [Li, Ri), Li = Ri-1, L1 = min(fp), RM = max(fp)+1, i.e., Si = { ( fp,fo ) | Li  fp < Ri } • Mi = COMPUTE MODEL for Si, i=1..M, where: Mi = • OUTPUT the complete model M = {M1,…, MM} ) , , , , , , , , p b a L b L a U b U a L b L a U b U a l l l l s s s s   ( ) , , , , , , , , p b a L b L a U b U a L b L a U b U a l l l l s s s s   ( The Enhanced VGA Partitioning the Dataset
  • 18. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 18 of 25 Experimental Results • Experiments have been done with an elitist generational GA/VGA (popsize = 100, pX = .9, pm=.01, maxevals=10000) and PGA/PVGA (popsize = 300/400, pX = .9, pm = .01, maxevals = = 20000) using ranking selection. • All results correspond to series of 50 runs. • Three problems have been considered: the Rastrigin function, the Fletcher-Powell function, and flowshop scheduling. • Issues tackled: – Goodness of quantitative estimation – Goodness of qualitative estimation – Computational speed-up Experimental Results
  • 19. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 19 of 25 Quantitative Estimations Fletcher-Powell Function Experimental Results Quantitative Estimations
  • 20. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 20 of 25 Quantitative Estimations Flowshop Scheduling Experimental Results Quantitative Estimations
  • 21. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 21 of 25 Qualitative Estimations Three different GA configurations are compared on the Rastrigin Function: A = {popsize=100, pX=.9, pm=.01} B = {popsize=50, pX=.5, pm=.10} C = {popsize=80, pX=.7, pm=.05} Real GA Virtual GA Relative properties are preserved. Experimental Results Qualitative Estimations
  • 22. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 22 of 25 Qualitative Estimations Four different Parallel GA (PGA) configurations are compared on a flowshop scheduling problem: A = {GAs=1}, C = {GAs=4} B = {GAs=2}, D = {GAs=8} Real PGA Virtual PGA Relative properties are preserved. Experimental Results Qualitative Estimations (ring topology)
  • 23. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 23 of 25 Qualitative Estimations Three different PGA configurations are compared on a flowshop scheduling problem: A = {GAs = 2, migfreq = 1000, migrate-best} B = {GAs = 2, migfreq = 300, migrate-random} C = {GAs = 4, migfreq = 500, migrate-random} Real PGA Virtual PGA Relative properties are preserved. Experimental Results Qualitative Estimations
  • 24. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 24 of 25 Speed-Up The larger the problem instance, the higher the speed-up of the VGA vs. GA. Experimental Results Speed-Up
  • 25. On Partitioned Fitness- Distributions of Genetic Operators for Predicting GA Performance Outline of the Presentation Introduction Virtual Genetic Algorithms Experimental Results Conclusions Sketch of a VGA The Basic VGA Problems of the Basic VGA Enhancing the Model Quantitative Estimations Qualitative Estimations Speed-Up Some Results Partitioning the Dataset The Enhanced VGA 9/26/2021 25 of 25 Conclusions • VGAs can be a useful tool for predicting GA performance. • An enhanced VGA model has been proposed to grasp some properties of the fitness distribution such as asymmetry, statistical anisotropy, or the different behaviour of GA. • The obtained results show: ✓ a much more accurate performance estimation ✓ high computational speed-ups • Future work: study other schemes for partitioning the dataset and other non-linear statistical models Conclusions