Multi-Population Adaptive Inflationary Differential Evolution - - PowerPoint PPT Presentation

▶
multi population adaptive inflationary differential
SMART_READER_LITE
LIVE PREVIEW

Multi-Population Adaptive Inflationary Differential Evolution - - PowerPoint PPT Presentation

Sep 27, 2022 379 likes •1.24k views

- - - - Multi-Population Adaptive Inflationary Differential Evolution Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Department of Mechanical and Aerospace Engineering University of Strathclyde marilena.di-carlo@strath.ac.uk PPSN

slide-1
SLIDE 1
  • Multi-Population Adaptive Inflationary

Differential Evolution

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci

Department of Mechanical and Aerospace Engineering University of Strathclyde marilena.di-carlo@strath.ac.uk

PPSN BIOMA - Bioinspired Optimization Methods and their Applications Ljubljana, 13 September 2014

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-2
SLIDE 2

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Introduction

◮ Differential Evolution (DE) is a very efficient population-based

stochastic algorithm for global numerical optimization problems

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-3
SLIDE 3

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Introduction

◮ Differential Evolution (DE) is a very efficient population-based

stochastic algorithm for global numerical optimization problems

◮ Its performance can be enhanced by combining it with others

  • ptimizer:

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-4
SLIDE 4

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Introduction

◮ Differential Evolution (DE) is a very efficient population-based

stochastic algorithm for global numerical optimization problems

◮ Its performance can be enhanced by combining it with others

  • ptimizer:

Inflationary Differential Evolution Algorithm, IDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-5
SLIDE 5

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Introduction

◮ Differential Evolution (DE) is a very efficient population-based

stochastic algorithm for global numerical optimization problems

◮ Its performance can be enhanced by combining it with others

  • ptimizer:

Inflationary Differential Evolution Algorithm, IDEA

◮ DE performance are strongly influenced by setting of the

algorithm parameter:

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-6
SLIDE 6

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Introduction

◮ Differential Evolution (DE) is a very efficient population-based

stochastic algorithm for global numerical optimization problems

◮ Its performance can be enhanced by combining it with others

  • ptimizer:

Inflationary Differential Evolution Algorithm, IDEA

◮ DE performance are strongly influenced by setting of the

algorithm parameter:

Adaptive Inflationary Differential Evolution Algorithm, AIDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-7
SLIDE 7

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Introduction

◮ Differential Evolution (DE) is a very efficient population-based

stochastic algorithm for global numerical optimization problems

◮ Its performance can be enhanced by combining it with others

  • ptimizer:

Inflationary Differential Evolution Algorithm, IDEA

◮ DE performance are strongly influenced by setting of the

algorithm parameter:

Adaptive Inflationary Differential Evolution Algorithm, AIDEA

◮ Multi-population version of AIDEA

(MP-AIDEA)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-8
SLIDE 8

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Contents

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-9
SLIDE 9

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Contents

  • Differential Evolution

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-10
SLIDE 10

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Contents

  • Differential Evolution
  • IDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-11
SLIDE 11

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Contents

  • Differential Evolution
  • IDEA
  • AIDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-12
SLIDE 12

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Contents

  • Differential Evolution
  • IDEA
  • AIDEA
  • Multi-Population AIDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-13
SLIDE 13

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Contents

  • Differential Evolution
  • IDEA
  • AIDEA
  • Multi-Population AIDEA
  • Test Results

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-14
SLIDE 14

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

◮ Initialize population in the search

space

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-15
SLIDE 15

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

◮ Initialize population in the search

space

◮ Select three individuals x1, x2 and x3

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-16
SLIDE 16

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

x2 x3 x1

◮ Initialize population in the search

space

◮ Select three individuals x1, x2 and x3

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-17
SLIDE 17

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

x3 x1 (x2 −x3) x2 (x2 −x3) F v1

◮ Initialize population in the search

space

◮ Select three individuals x1, x2 and x3 ◮ Apply mutation: v1 = x1 + F · (x2 − x3)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-18
SLIDE 18

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

x1 v1 u1

◮ Initialize population in the search

space

◮ Select three individuals x1, x2 and x3 ◮ Apply mutation: v1 = x1 + F · (x2 − x3) ◮ Apply crossover to obtain trial vector

u1:

uj

1 =

  • vj

1,

if rand(0,1) ≤ CR or j = jrand xj

1,

  • therwise

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-19
SLIDE 19

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

◮ Initialize population in the search

space

◮ Select three individuals x1, x2 and x3 ◮ Apply mutation: v1 = x1 + F · (x2 − x3) ◮ Apply crossover to obtain trial vector

u1:

uj

1 =

  • vj

1,

if rand(0,1) ≤ CR or j = jrand xj

1,

  • therwise

◮ Repeat operation for all the individuals

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-20
SLIDE 20

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

Differential Evolution

1 2 3 4 5 1 2 3 4 5 6

◮ Initialize population in the search

space

◮ Select three individuals x1, x2 and x3 ◮ Apply mutation: v1 = x1 + F · (x2 − x3) ◮ Apply crossover to obtain trial vector

u1:

uj

1 =

  • vj

1,

if rand(0,1) ≤ CR or j = jrand xj

1,

  • therwise

◮ Repeat operation for all the individuals ◮ Survival selection: x′

i =

  • ui,

if f (ui) ≤ f (xi) xi,

  • therwise

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-21
SLIDE 21

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

◮ DE drawbacks:

  • Stagnation of the optimization process
  • CR and F difficult to tune and heavily problem dependent

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-22
SLIDE 22

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

◮ DE drawbacks:

  • Stagnation of the optimization process
  • CR and F difficult to tune and heavily problem dependent

◮ IDEA (Inflationary Differential Evolution Algorithm)

  • M. Vasile, E. Minisci, M. Locatelli, 2011
  • Hybridization of DE with the restarting procedure of Monotonic

Basin Hopping (MBH) algorithm

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-23
SLIDE 23

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 1. Initialize population in the search space and run Differential

Evolution

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-24
SLIDE 24

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 1. Initialize population in the search space and run Differential

Evolution

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-25
SLIDE 25

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 2. Population contraction: r ≤ ρ · rmax

r = max (xi − xj) and rmax is the maximum value of r recorded during the convergence

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-26
SLIDE 26

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 3. Perform local search from the best individual in the population and

locate local minimum

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-27
SLIDE 27

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 4. Restart population in a bubble of dimension δlocal around the

local minimum

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-28
SLIDE 28

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 5. Repeat DE until convergence

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-29
SLIDE 29

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 5. Repeat DE until convergence

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-30
SLIDE 30

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 5. Repeat DE until convergence

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-31
SLIDE 31

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

IDEA

  • 6. When more than nLR local restarts have been performed globally

restart the population at a distance δglobal from the centers of the clusters of local minima

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-32
SLIDE 32

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

AIDEA

◮ DE drawbacks:

  • Stagnation of the optimization process
  • CR and F difficult to tune and heavily problem dependent

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-33
SLIDE 33

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

AIDEA

◮ DE drawbacks:

  • Stagnation of the optimization process
  • CR and F difficult to tune and heavily problem dependent

◮ AIDEA (Adaptive Inflationary Differential Evolution Algorithm)

  • E. Minisci, M. Vasile, 2014
  • Hybridization of DE with the restaring procedure of Monotonic

Basin Hopping (MBH) algorithm

  • Adaptation of the DE parameters CR and F

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-34
SLIDE 34

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

AIDEA

CR and F adaptation

◮ Initialization of CRF, regular mesh in the space:

CR ∈ [0.1, 0.99] F ∈ [−0.5, 1]

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-35
SLIDE 35

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

AIDEA

CR and F adaptation

◮ Initialization of CRF, regular mesh in the space:

CR ∈ [0.1, 0.99] F ∈ [−0.5, 1]

◮ Sample of CR and F values from the Parzen distribution associated

to CRF and association of each (CR, F) couple to an individual of the population

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-36
SLIDE 36

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

AIDEA

CR and F adaptation

◮ Initialization of CRF, regular mesh in the space:

CR ∈ [0.1, 0.99] F ∈ [−0.5, 1]

◮ Sample of CR and F values from the Parzen distribution associated

to CRF and association of each (CR, F) couple to an individual of the population

◮ Computation for each individual x and its child x′ of the difference:

d = f (x′) − f (x)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-37
SLIDE 37

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Differential Evolution IDEA AIDEA

AIDEA

CR and F adaptation

◮ Initialization of CRF, regular mesh in the space:

CR ∈ [0.1, 0.99] F ∈ [−0.5, 1]

◮ Sample of CR and F values from the Parzen distribution associated

to CRF and association of each (CR, F) couple to an individual of the population

◮ Computation for each individual x and its child x′ of the difference:

d = f (x′) − f (x)

◮ CRF update: (CR, F) couples with lower d are substituted by

(CR, F) couples with higher d

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-38
SLIDE 38

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-39
SLIDE 39

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

◮ Adaptation of other parameters: multi-population AIDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-40
SLIDE 40

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

◮ Adaptation of other parameters: multi-population AIDEA ◮ Adaptation of the dimension of the bubble for the local restart δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-41
SLIDE 41

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

◮ Adaptation of other parameters: multi-population AIDEA ◮ Adaptation of the dimension of the bubble for the local restart δlocal ◮ Strategies for the generation of the mutant vector:

  • DE/best-DE/rand
  • DE/arch-DE/rand
  • DE/arch-DE/best

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-42
SLIDE 42

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

◮ Adaptation of other parameters: multi-population AIDEA ◮ Adaptation of the dimension of the bubble for the local restart δlocal ◮ Strategies for the generation of the mutant vector:

  • DE/best-DE/rand
  • DE/arch-DE/rand
  • DE/arch-DE/best

◮ CR and F adaptation:

  • MP-AIDEA-CRF1

Same CR and F values for every individual of the same population

  • MP-AIDEA-CRF2

Different CR and F values for each element of each population

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-43
SLIDE 43

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

δlocal adaptation

◮ Computation of the minimum and maximum distance between

all local minima, dminMIN and dminMAX

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-44
SLIDE 44

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

δlocal adaptation

◮ Computation of the minimum and maximum distance between

all local minima, dminMIN and dminMAX

◮ Creation of a regular mesh B in the space [dminMIN, dminMAX]

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-45
SLIDE 45

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

δlocal adaptation

◮ Computation of the minimum and maximum distance between

all local minima, dminMIN and dminMAX

◮ Creation of a regular mesh B in the space [dminMIN, dminMAX] ◮ Sample of δlocal from the Parzen distribution associated to B

for each population

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-46
SLIDE 46

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

δlocal adaptation

◮ Computation of the minimum and maximum distance between

all local minima, dminMIN and dminMAX

◮ Creation of a regular mesh B in the space [dminMIN, dminMAX] ◮ Sample of δlocal from the Parzen distribution associated to B

for each population

◮ Computation for each population of the distance between

consecutive local minima: p =

  • xk+1

min − xk min

  • Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci

Multi-Population Adaptive Inflationary Differential Evolution

slide-47
SLIDE 47

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA

δlocal adaptation

◮ Computation of the minimum and maximum distance between

all local minima, dminMIN and dminMAX

◮ Creation of a regular mesh B in the space [dminMIN, dminMAX] ◮ Sample of δlocal from the Parzen distribution associated to B

for each population

◮ Computation for each population of the distance between

consecutive local minima: p =

  • xk+1

min − xk min

  • ◮ Update of B: population with higher values of p are

characterized by a better value of δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-48
SLIDE 48

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA versions

CRF1 CRF2 DE-mut1 DE-mut2 DE-mut3 δlocal MP-AIDEA 1

  • MP-AIDEA 2
  • MP-AIDEA 2
  • MP-AIDEA 4
  • MP-AIDEA 5
  • MP-AIDEA 6
  • MP-AIDEA 7
  • MP-AIDEA 8
  • MP-AIDEA 9
  • MP-AIDEA 10
  • MP-AIDEA 11
  • MP-AIDEA 12
  • DE-mut1: DE/best-DE/rand

DE-mut2: DE/arch-DE/rand DE-mut3: DE/arch-DE/best Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-49
SLIDE 49

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions MP-AIDEA δlocal adaptation MP-AIDEA versions

MP-AIDEA versions

CRF1 CRF2 DE-mut1 DE-mut2 DE-mut3 δlocal MP-AIDEA 1

  • MP-AIDEA 2
  • MP-AIDEA 2
  • MP-AIDEA 4
  • MP-AIDEA 5
  • MP-AIDEA 6
  • MP-AIDEA 7
  • MP-AIDEA 8
  • MP-AIDEA 9
  • MP-AIDEA 10
  • MP-AIDEA 11
  • MP-AIDEA 12
  • DE-mut1: DE/best-DE/rand

DE-mut2: DE/arch-DE/rand DE-mut3: DE/arch-DE/best Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-50
SLIDE 50

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Test Cases

Competition of the Congress on Evolutionary Computation (CEC)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-51
SLIDE 51

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Test Cases

Competition of the Congress on Evolutionary Computation (CEC)

  • 1. Spread Spectrum Radar Polyphase Code Design, CEC 2011
  • 2. Tersoff Radar Function Minimization Problem, CEC 2011
  • 3. Schwefel’s Problem, CEC 2005
  • 4. Rotated Version of Hybrid Composition Function, CEC 2005

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-52
SLIDE 52

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Test Cases

Competition of the Congress on Evolutionary Computation (CEC)

  • 1. Spread Spectrum Radar Polyphase Code Design, CEC 2011
  • 2. Tersoff Radar Function Minimization Problem, CEC 2011
  • 3. Schwefel’s Problem, CEC 2005
  • 4. Rotated Version of Hybrid Composition Function, CEC 2005

Algorithm performance:

  • Success rate: number of successful runs over 100 total runs
  • Successful run: f (xmin) < fmin + ǫ

CEC 2011: ǫ = 0.001 CEC 2005: ǫ = 0.01

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-53
SLIDE 53

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Spread Spectrum Radar Polyphase Code Design

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-54
SLIDE 54

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Spread Spectrum Radar Polyphase Code Design

◮ Problem and algorithm parameters

D fmin FEs δlocal* δglobal ρ nLR 20 0.5 1.5·105 0.1 0.1 0.2 10 * for AIDEA and MP-AIDEA versions which do not adapt δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-55
SLIDE 55

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Spread Spectrum Radar Polyphase Code Design

◮ Problem and algorithm parameters

D fmin FEs δlocal* δglobal ρ nLR 20 0.5 1.5·105 0.1 0.1 0.2 10 * for AIDEA and MP-AIDEA versions which do not adapt δlocal

◮ Algorithm comparison

  • AIDEA
  • Best performing algorithms of CEC 2011 competition:

GA-MPC (Genetic Algorithm with Multi-Parent Crossover) WI-DE (Weed Inspired Differential Evolution)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-56
SLIDE 56

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Spread Spectrum Radar Polyphase Code Design

◮ MP-AIDEA and AIDEA success rate

2 4 6 8 10 12 14 16 18 20 22 20 40 60 80 100

Npop Success rate MP−AIDEA DE/best−DE/rand

AIDEA MP−AIDEA 1 MP−AIDEA 4 MP−AIDEA 7 MP−AIDEA 10

2 4 6 8 10 12 14 16 18 20 22 20 40 60 80 100

Npop Success rate MP−AIDEA DE/arch−DE/rand

AIDEA MP−AIDEA 2 MP−AIDEA 5 MP−AIDEA 8 MP−AIDEA 11

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-57
SLIDE 57

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Spread Spectrum Radar Polyphase Code Design

◮ Statistics of the results Algorithm Min Mean Max Str.Dev. MP-AIDEA 10 0.5000 0.5045 0.5690 0.0135 AIDEA 0.5000 0.5130 0.6422 0.0263 GA-MPC 0.5000 0.7484 0.9334 0.1249 WI-DE 0.5000 0.6560 0.9931 0.1160

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-58
SLIDE 58

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Tersoff Potential Function Minimization Problem

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-59
SLIDE 59

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Tersoff Potential Function Minimization Problem

◮ Problem and algorithm parameters D fmin FEs 30

  • 36.9286

1.5·105 δlocal* δglobal ρ nLR Case 1 0.1 0.1 0.2 10 Case 2 0.3 0.1 0.2 10

* for AIDEA and MP-AIDEA versions which do not adapt δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-60
SLIDE 60

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Tersoff Potential Function Minimization Problem

◮ Problem and algorithm parameters D fmin FEs 30

  • 36.9286

1.5·105 δlocal* δglobal ρ nLR Case 1 0.1 0.1 0.2 10 Case 2 0.3 0.1 0.2 10

* for AIDEA and MP-AIDEA versions which do not adapt δlocal

◮ Algorithm comparison

  • AIDEA
  • Best performing algorithms of CEC 2011 competition:

GA-MPC (Genetic Algorithm with Multi-Parent Crossover) WI-DE (Weed Inspired Differential Evolution)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-61
SLIDE 61

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Tersoff Potential Function Minimization Problem

◮ MP-AIDEA and AIDEA success rate

2x10 4x10 6x10 8x10 20 40 60 Npopxnpop Success rate Case 1

AIDEA MP−AIDEA 1 MP−AIDEA 4 MP−AIDEA 7 MP−AIDEA 10

2x10 4x10 6x10 8x10 20 40 60 Npopxnpop Success rate Case 2

AIDEA MP−AIDEA 1 MP−AIDEA 4 MP−AIDEA 7 MP−AIDEA 10

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-62
SLIDE 62

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Tersoff Potential Function Minimization Problem

◮ δlocal adaptation 10 20 30 0.05 0.1 0.15 0.2 0.25 0.3 0.35 Adaptation Steps δlocal Population1 Population2 Population3 MeanValue

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-63
SLIDE 63

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Tersoff Potential Function Minimization Problem

◮ Statistics of the results Algorithm Min Mean Max Str.Dev. Case 1 MP-AIDEA 7

  • 36.9286
  • 36.7120
  • 34.3504

0.3835 AIDEA

  • 36.9286
  • 36.8046
  • 35.9700

0.2483 Case 2 MP-AIDEA 10

  • 36.9286
  • 36.6689
  • 34.1647

0.4399 AIDEA

  • 36.9286
  • 36.6219
  • 35.4467

0.4694 GA-MPC

  • 36.8457
  • 35.03883
  • 34.1076

0.8329 WI-DE

  • 36.8
  • 35.6
  • 34.2

0.904

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-64
SLIDE 64

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-65
SLIDE 65

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

◮ Problem and algorithm parameters

D fmin FEs δlocal* δglobal ρ nLR 30/50

  • 460

3·105 / 5·105 0.1 0.1 0.2 5 * for AIDEA and MP-AIDEA versions which do not adapt δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-66
SLIDE 66

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

◮ Problem and algorithm parameters

D fmin FEs δlocal* δglobal ρ nLR 30/50

  • 460

3·105 / 5·105 0.1 0.1 0.2 5 * for AIDEA and MP-AIDEA versions which do not adapt δlocal

◮ Algorithm comparison

  • AIDEA
  • Best performing algorithms of CEC 2005 competition:

IPOP-CMA-ES (Increasing Population Size Covariance Matrix Adaptation Evolution Strategy)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-67
SLIDE 67

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

◮ MP-AIDEA and AIDEA success rate

2 4 6 8 10 12 14 16 20 40 60 80 100 Npop

Success rate

MP−AIDEA DE/best−DE/rand

AIDEA MP−AIDEA 1 MP−AIDEA 4 MP−AIDEA 7 MP−AIDEA 10

D = 30 npop = 10

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-68
SLIDE 68

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

◮ MP-AIDEA and AIDEA success rate

2 4 6 8 10 12 14 16 20 40 60 80 100 Npop

Success rate

MP−AIDEA DE/best−DE/rand

AIDEA MP−AIDEA 1 MP−AIDEA 4 MP−AIDEA 7 MP−AIDEA 10

D = 30 npop = 10

2 4 6 8 5 10 15

Npop Success rate

MP−AIDEA DE/best−DE/rand AIDEA MP−AIDEA 1 MP−AIDEA 4 MP−AIDEA 7 MP−AIDEA 10

D = 50 npop = 20

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-69
SLIDE 69

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

◮ Statistics of the results (error values w.r.t. fmin) Algorithm Min Mean Max Str.Dev. D = 30 MP-AIDEA 10 1.39e-9 2.45e+1 4.77e+2 7.25e+1 AIDEA 2.01e-9 1.03e+2 1.00e+3 1.97e+2 IPOP-CMA-ES 3.79e-9 4.43e+4 1.10e+6 2.19e+5 D = 50 MP-AIDEA 10 2.48e-8 8.91e+2 9.54e+3 1.27e+3 AIDEA 5.61e-8 2.22e+3 1.33e+4 2.69e+3 IPOP-CMA-ES 9.67e+0 2.27e+5 5.57e+6 1.11e+6

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-70
SLIDE 70

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Schwefel’s Problem

◮ Statistics of the results (error values w.r.t. fmin) Algorithm Min Mean Max Str.Dev. D = 30 MP-AIDEA 10 1.39e-9 2.45e+1 4.77e+2 7.25e+1 AIDEA 2.01e-9 1.03e+2 1.00e+3 1.97e+2 IPOP-CMA-ES 3.79e-9 4.43e+4 1.10e+6 2.19e+5 D = 50 MP-AIDEA 10 2.48e-8 8.91e+2 9.54e+3 1.27e+3 AIDEA 5.61e-8 2.22e+3 1.33e+4 2.69e+3 IPOP-CMA-ES 9.67e+0 2.27e+5 5.57e+6 1.11e+6

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-71
SLIDE 71

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Rotated Version of Hybrid Composition Function

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-72
SLIDE 72

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Rotated Version of Hybrid Composition Function

◮ Problem and algorithm parameters

D fmin FEs δlocal* δglobal ρ nLR 10 120 1·105 0.1 0.1 0.2 5 * for AIDEA and MP-AIDEA versions which do not adapt δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-73
SLIDE 73

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Rotated Version of Hybrid Composition Function

◮ Problem and algorithm parameters

D fmin FEs δlocal* δglobal ρ nLR 10 120 1·105 0.1 0.1 0.2 5 * for AIDEA and MP-AIDEA versions which do not adapt δlocal

◮ Algorithm comparison

  • AIDEA
  • Best performing algorithms of CEC 2005 competition:

IPOP-CMA-ES (Increasing Population Size Covariance Matrix Adaptation Evolution Strategy)

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-74
SLIDE 74

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Rotated Version of Hybrid Composition Function

◮ MP-AIDEA success rate

Algorithm 2x20 4x20 6x20 8x20 MP-AIDEA 4 2 1 1 MP-AIDEA 10 1 1 3

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-75
SLIDE 75

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions Test Cases Spread Spectrum Radar Polyphase Code Design Tersoff Potential Function Minimization Problem Schwefel’s Problem Rotated Version of Hybrid Composition Function

Rotated Version of Hybrid Composition Function

◮ MP-AIDEA success rate

Algorithm 2x20 4x20 6x20 8x20 MP-AIDEA 4 2 1 1 MP-AIDEA 10 1 1 3

◮ Statistics of the results (error values w.r.t. fmin)

Algorithm Min Mean Max Str.Dev. MP-AIDEA 10 7.44e-11 1.05e+2 1.32e+2 2.35e+1 AIDEA 5.38e+1 1.02e+2 1.14e+2 8.42e+0 IPOP-CMA-ES 7.92e+1 9.13e+1 9.68e+1 3.49e+0

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-76
SLIDE 76

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Conclusions

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-77
SLIDE 77

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Conclusions

◮ Multi-population version of AIDEA

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-78
SLIDE 78

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Conclusions

◮ Multi-population version of AIDEA

  • Hybridization of DE and MBH
  • Adaptation of CR, F and δlocal

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-79
SLIDE 79

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Conclusions

◮ Multi-population version of AIDEA

  • Hybridization of DE and MBH
  • Adaptation of CR, F and δlocal

◮ Test results:

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-80
SLIDE 80

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Conclusions

◮ Multi-population version of AIDEA

  • Hybridization of DE and MBH
  • Adaptation of CR, F and δlocal

◮ Test results:

  • MP-AIDEA with adaptation of δlocal give good results and do

not require the setting of this parameter

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-81
SLIDE 81

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Conclusions

◮ Multi-population version of AIDEA

  • Hybridization of DE and MBH
  • Adaptation of CR, F and δlocal

◮ Test results:

  • MP-AIDEA with adaptation of δlocal give good results and do

not require the setting of this parameter

  • Most successful versions of MP-AIDEA were able to

locate for the first time the global minima of two difficult academic test functions

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-82
SLIDE 82

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Future work

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-83
SLIDE 83

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Future work

◮ Adaptation of other parameters:

maximum number of local restart nLR

2x10 4x10 6x10 8x10 2x10 4x10 6x10 10 20 30 40 50 60 Npopxnpop Success rate Tersoff Potential Minimization Problem MP−AIDEA 4 NEW MP−AIDEA 4 MP−AIDEA 10 NEW MP−AIDEA 10 Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution

slide-84
SLIDE 84

Introduction IDEA & AIDEA Multi-Population AIDEA Test Results Conclusions

Thank you for your attention

Marilena Di Carlo, Massimiliano Vasile, Edmondo Minisci Multi-Population Adaptive Inflationary Differential Evolution