10
Interactive Genetic Algorithm using Initial Individuals Generated from Human Sensitivity Noriko OKADA * Mitsunori MIKI ** Tomoyuki HIROYASU *** and Masato YOSHIMI ** (Received January 17, 2011) In this paper, we propose the new methods to generate initial individual which reflects human’s sensitivity in Interactive Genetic Algorithm (IGA). Specifically, we propose the initial individuals generation method based on color harmony theories. IGA is an optimization method based on Genetic Algorithms (GA) which simulates the evolution of living things, where the evaluation part of the GA is handled subjectively by a user. Color harmony theory are the principles used to create harmonious color combinations. In the proposed methods, by including user’s favorite individuals in an initial population, we aim at to increase efficiency of searching solution and reducing user’s loads. We constructed a system which designs a color combination of individual workspace and experimented to verify the validity of the proposal methods. The experiment showed that a design with a user’s high level of satisfaction is generable in the system using the proposal methods. In addition, we figured out that the proposed methods are effective, and found out that it was useful in reducing the psychological fatigue of the users. Key words optimization, interactive evoluationary method, Interactive Genetic Algorithm, color combination 1. 1) * Graduate Student, Department of Intelligent Information Engineering and Sciences, Doshisha University, Kyoto Telephone:+81-774-65-6924, E-mail:[email protected] ** Department of Intelligent Information Engineering and Sciences, Doshisha University, Kyoto Telephone:+81-774-65-6930, Fax:+81-774-65-6796, E-mail:{mmiki, myoshimi}@mail.doshisha.ac.jp *** Department of Biomedical Information, Doshisha University, Kyoto Telephone:+81-774-65-6932, Fax:+81-774-65-6019, E-mail:[email protected] 2) (Interactive Genetic Algorithm:IGA) 3)

Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

Interactive Genetic Algorithm using Initial Individuals Generated

from Human Sensitivity

Noriko OKADA* Mitsunori MIKI** Tomoyuki HIROYASU*** and Masato YOSHIMI**

(Received January 17, 2011)

In this paper, we propose the new methods to generate initial individual which reflects human’s sensitivity in Interactive

Genetic Algorithm (IGA). Specifically, we propose the initial individuals generation method based on color harmony theories.

IGA is an optimization method based on Genetic Algorithms (GA) which simulates the evolution of living things, where

the evaluation part of the GA is handled subjectively by a user. Color harmony theory are the principles used to create

harmonious color combinations. In the proposed methods, by including user’s favorite individuals in an initial population, we

aim at to increase efficiency of searching solution and reducing user’s loads. We constructed a system which designs a color

combination of individual workspace and experimented to verify the validity of the proposal methods. The experiment showed

that a design with a user’s high level of satisfaction is generable in the system using the proposal methods. In addition, we

figured out that the proposed methods are effective, and found out that it was useful in reducing the psychological fatigue of

the users.

Key words �optimization, interactive evoluationary method, Interactive Genetic Algorithm, color combination����� �������� ��������������������� !"#$%&'()*+,-./0123456789:;<=>? @A B CD E F G CH I J K CL M N O

1. PQRSTU�VW�X YZ[\]^_`a�bc�de^fgahijkl�mfnopqrstj�de�uvjwqxya`s1) zX YZ[\]{|}p~�s�^���vk������bc^�yauvk����sw����k��Y��p��k`�w�Z[\]Y* Graduate Student, Department of Intelligent Information Engineering and Sciences, Doshisha University, Kyoto

Telephone:+81-774-65-6924, E-mail:[email protected]

** Department of Intelligent Information Engineering and Sciences, Doshisha University, Kyoto

Telephone:+81-774-65-6930, Fax:+81-774-65-6796, E-mail:{mmiki, myoshimi}@mail.doshisha.ac.jp

*** Department of Biomedical Information, Doshisha University, Kyoto

Telephone:+81-774-65-6932, Fax:+81-774-65-6019, E-mail:[email protected]

X p� ���{���{k`z���r�X YZ[\]����Y��w�v�ks 2) z���� �tjY¡�{¢�^£yadew¤ks�`x`k¥���s�r¦§��_£¨|Z��w©ª��sz�������«^� ��¥�p��¬��� �­n^®¯°]±²³´^���µ�s¶���a�����������(Interactive Genetic Algorithm:IGA) 3)

Page 2: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

����������� �����������������w� µ!a`szIGA{���������� "

Genetic Algorithm:

GA#4) ^£s$%p&³'���� w�~¡�Y()kl�tjp­n*+��a,rs-p.¯/0¶���szIGA�{-$%^� �*1w�2��s�r�¢3+Y$%45+p670s��^£�8³[�9:p;<0s�vw�sz����¢3+Y$%45+p67{=>?@�`�A�^~kwszBC�k IGA�{X D+E FG�HI]J�^KL¢3pMN0sz����IGA

�{�O�^MNµ!s¢3wKL¢3^PQ0sRSw�s�r�8³[�T¡^��k`¢3wKL¢3U^V�Wx!a`sX�^{�8³[wYZ,rs-^?@���[\-^]�Y0`�`�A�¥�szµH^�I]J�kKL¢3MN^£�T¡^��k`¢3wV�̂ _µ!s��8³[^­np���^`a�bcp1gs��¥;gH!sz����Yde�{IGA�-$%^_fs8³[�9:pgh0s��p ����IGA

^_fsKL¢3MNi�^j 0szk�A�p)����lmnopp`aKL¢3pMN0s¶�p̂ q0sz�lmnopp`�8³[�T¡^��¢3pKL¢3U^Wrs��^£��-$%���8³[�­nr+phH��9:gh�stwLu�¯sz2. vwxyz{|}~���

2.1 vwxyz{|}~������IGA�{�M����p|Z�������¶���s

GA^_fs�����p&³'���­n��^� ���^®�`�­npp`s����-$%p� �¶���sz� ^£s���­npp`a-$%p���r�

IGA{� �tjp�'��^��¬����`gs5) z���r��Z�GA

���a��tjp������^��a`s�`�!a`sz��£�k���H�IGA{¦§�k­nw©ªkZ[\]����_£¨�]´³�\�]�kl���k������pµ!a`s6, 7, 8) zYde�{�8³[�¡�p���

�� ¡�kT¡p¢¡�µ�s��p ����IGAp�p0szIGA�'���£¤¥pFig. 1

^_0zx��IGA^_fs®Y���¦§³¨©³�pFig. 2

^_0zUser System

GA

Display

Evaluation

Fig. 1. IGA system.

Initialization

Evaluation

Selection

Crossover

Mutation

Start

EndYes

No

Display

Human operation

Terminal

criterion

Fig. 2. Flow chart of IGA.

2.2 vwxyz{|}~���Sª«¬­®¯°�IGA^_fs-�$%E {�X E FG��szX E w��±²kX����-^?@0s�r^{V��$%45+�¢3w�v^ksz����

IGA�{��³´pV�0s�¢3p­n0s8³[^µ¯k9:p1ga�x�z8³[�9:pgh0s�r{³´p¶k��kf!·kHk`w��!{=>?@^~kwszx��BC�k IGA

�{I]J�^KL¢3wMNµ

Page 3: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

���� C�� �� C� � C�� �!s�r�KL$%�G{X D+E FG��sBi��^_���G�7�^£�̂ _¢3+{67p�fszIGA�{��-wKL¢3^PQ0sRSw���8³[�T¡^��¥�wKL¢3U^Wx!k`X�^{$%45+w�f�a�x�zIGA

�{^_µ!sFa�¢3^�a8³[w­np���vw�s�r�$%45+��f^�¥k`8³[^0sbc¥�f0sz¯°� w­n�¯s45+{ 10�H

2045�e�`�!a_��� w­n�¯sr+^¥7�w�sz���r�7H!�^_¢3+�s�£�$%p���vw�sz����KL¢3^�H��r8³[�T¡^��¢3pWrs���$%E p�r�-$%�s��p¥s��p;gszKL¢3^T¡j�q`¢3pWrsi���a�Yde�{�lmnopp`sz

3. ������� !"#$%&'3.1 �����3.1.1 ���������lmno�{�

2�()�Hks�������^�a� wt�s*C+*�i�`C,`-��lmn^_fsmnC+mn�./��s`{�/j^~`ade��3«�f�¥���s9) z0³§12�{�34�^56µ!a¯��lmno^®�`a�789]�:I³k�Y��°§³�kl���kk�w��!�;¡�¥\]���Y<=klV��Z[\]���>pµ!�¦j�a`sz�lmno{�?]@�A�«Y

PCCS(Practical Color Co-ordinate System)A�«

10)kl�`�!���l3«^;gH!a_��B�l3«C�^mnC+mn��DE�FG*HwI¦µ!a`szYde�{�k�mnp;gs�^���PCCSA�«pp`aAµ!s�lmnopp`sz3.1.2 PCCSJ�K

PCCS(Practical Color Co-ordinate System)A�«{�LY�lde\w�lmn�A�p�'�?�M1N^-O0s��p�k ���aPQ��1964U^QA��:I³�'����szV �kRpjw���k�

mn��S¦p�`Y0`�r��l ���lmT�U �_£¨�lVWkl��`�X���kpY^p`H!a`s10) zPCCSA�«�{�

HSBA�«�D�^���A;^�F�le�_£¨Ze�3[jpp`sw�µH^�F��³] (tone)

�`�2~�v\�H�p�]0s^w�µ�_`��szPCCS��³]�{�

Fig. 3^_0£�̂ le_£¨Ze�±�£¤����Zs`�a`�b`�`y���m�pA0¥���szPCCS��F�{�`acd���se�f�g�_£¨hp�v�F��a`szPCCS

��Fi�{��!H4�^�a��`aj�pSkl^¦r�µH^�F mw-��ks£�̂ 4

�pfg�12�F^n����Fp�o�a 24

�F^kya`szPCCS��FipFig. 4^_0z

pPale

ltgLight

Grayish

g Grayish

vVivid

bBright

sStrong

dpDeep

ltLight

sfSoft

dk Dark

dDull

dkgDark

Grayish

WWhite

ltGyLight Gray

mGyMedium Gray

dkGyDark Gray

BkBlack

Saturation HighLow

Brig

htn

ess

Low

High

Fig. 3. Tone of PCCS.

3.2 p�qrSst¬uv�lmno^_fsk���{��Q¡0sz����l�k���wIGA�KL¢3MN^p`s�^wµ��`�pxy0s�r�20z5�{| 20}p~�´��a��R�p�y�zR��{�~�´^¡¯k�p

1��N�a¥Hy������p¥�^��kk���p�p�a�N�� 3

�k�p���Bk��¡�^~`ar��a¥Hy�zR���^�p��k��

Page 4: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

����������� �����������������2:R

3:yR

4:rO

5:O

6:yO

7:rY8:Y

9:gY

10:YG

11:yG

13:bG

14:BG

12:G

15:BG

16:gB

17:B

21:bP

24:RP

23:rP

22:P

20:V19:pB

18:B

1:pR

� �������� ����yellowish red

� ����reddish orange

������ ������ �� ���� !red

yello

wish

gree

n

green

blu

ish p

urp

le

redd

ish

purp

lepurplish red

yello

w g

reen

red purple

purp

lish b

lue

blue green

blue

bluish green

purp

le

vio

let

"#$$%&'( #))*+greenish blue

Fig. 4. Hue circle of PCCS.�{��³]C,]C�³]k���³]C\]C�³]k��¦-:?\8k��'.�1�C°].��]´�³k���I\�7k��/IZ³�0]k��9¨²I�1³|2³k��_£¨°]. 31N'1³|2³k�� 8 4]��s10) z�!H{�k����5�¥®Y����k���p`H!a`s5A�kk��_£¨678 9kl��:�C;<�^p`H!a`sk���=�]>��¥�p?`�¥��kya`szk_�R�^_fsk�����p@{I]J����zR��At��³]C,]C�³]k��¦-:?\8k��/IZ³�0]k��_£¨9¨²I�1³|2³k�{�BC`kk�D�r���~�´wF3� 3��1(E�����F�¡x!sRS^�s��w��y�zBi��³]C\]C�³]k��'.�1�C°].��]´�³k���I\�7k��_£¨°]. 31N'1³|2³k�{¡x!k`RS^�s��w��y�z�!H����H�Yde�{¡x!sRS^�y� 4 4�k���p�lmnao^®��KL¢3MNi�^�p0sz�p0sBk���^~`a(E^GZ����k�Hp

Fig. 5^_0z

�³]C,]C�³]k�DB�F��³]�ZeIp�F�µ¯��y�k

���sz•

¦-:?\8k��F�_£¨�³]^YYD�p~f��������^£sk���sz•

/IZ³�0]k��l�Jm��skK^£�L M�stp1gs3�()�k���sz/IZ³�0]k�^{���[j^N�a�F�/IZ³�0]�Ze�/IZ³�0]�le�/IZ³�0]�_£¨ �³]�/IZ³�0]�

4¯�w�sz

9¨²I�1³|2³k�k�pON0s���=��Fi^_fsf^T`�F��p�F�Zs���P^hQ^T`�F��pF�^R���k���szFundamental color

Faux-CamaieuTone on Tone

Gradation Natural harmony

Fig. 5. Example of color combination.

3.3 �����SST U"#$%&'3.2 V�W����R��At£��Y^q¶��{k�����a �³]C,]C�³]k��¦-:?\8k��/IZ³�0]k��_£¨9¨²I�1³|2³k�� 4 4pp`sz8³[{�'��pp`a¡¯k�p

1��X^�N0sz�'��{�8³[^£��Nµ!��p¥�^)Y

4 4]�k���p�p��8³[�N�pW� 3�k�p�N0sz��a�k��B�pKL¢3�2³�M�0]�Z'N�_£¨}³�2Z°][\^p`sz�[�]

3\[�=l[\wo �̂ aH!s�{I]J_^O¦0szx�]k\�N[�^`pµ!sk\�a^*�a¥�'�_wB�

Page 5: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

���� C�� �� C� � C�� �I]J_^O¦0szk\�app`�\lmno^®��KL¢3MNiapFig. 6^_0z

Fig. 6. Method of generating initial individuals based

on color the harmony theory.

4. ���r������!R�����^q¶a[�sjpx 0s�r[�'�_��a],¦M'^_fs¢�[��E [k\Z[\]p� �,¦M'E Z[\]�'�_pO���z(E^O����'�_[��p_0z4.1 �����J��r•

ONv\2³�M�0]]Z'N]_£¨}³ �2Z°][3~[�\�_�Z[\]pON0sz�[

3~[�\�_[\pD�0s��^£�]¢�[��E pA;0sz

\[A;iaY�'�_[\[A;ia^{]HSB

A\« 10)pp`szHSB

A\«{]� [tj^>�\[A;ia���]\F(Hue)

]le(Saturation)

]_£¨Ze(Brightness)

[3N�^£�\pA;0sz\F�{]e]h]_£¨gkl[\[4][�����]e]f]g]h]_£¨Q[

5\Fp��)^- m^���\Fi�A0��w�¯sz\Fi�Aµ!s��^£�]h�f[ 5 \{g�`y�� [\ltM^>�A;w���kszle�{\[ !"#$%&'()[*+,-.]/'01)23 !"4\5]6)23789\+4

:;<'+=\[<>$%&'()[*+,-.]<'01)23<:[email protected]\()54.]6)23>?4.\()0B".5??4:;4C]\D=0"E

360[FG]/'CHI<'= 0

"E100JFG,K$+:;

• LMNOP1 QJLMN, 1 QJRSTU$%V&:;LMNJWXYZ5=[W�T\]JMJHSB

KJ^K_`$ab&:;4C[cdef\]5C)g[HSB%Mh5Ci:/'CHI<'JK=jklA[0.0"E1.0Jm^K$n)9;

4.2 opqrstuvwxywz{|}~�1. ���NJ��CHIc�

3.3�,��9����5��)g���N$��&:;��[ef\]=M$��&:9�J��$��S5c�&:;��S=c���5C)g[��J�@4M$�M��&:;�J�[ef\]=��S0��A9M5�M��$�nA[3M�M$��&:;�Ag[��A9�M5n)E�g):WM$��\ e¡U[Rf¢[CHI£�¤�¥¦UJM5�nAg[��S5���N+Agc�&:;4C[§ef\]5Ci:�¨©-9.J�N^=[��ª5c�,@:�N^5=«¬0-:9�9+A9;

2. ­®��S=c�#�9�N 1Q 1Q5¯A[°±5��)g1"E

5²J 5³´,­®$µ¶;**,=[5 ²0·1²+A9;�9[W¨©,c�#�:9�JRSTUJ¶̧ [¹5º5»89¼J=½f¤RSTU+Ag¾¿&:;*�=[ÀÁ,Â<&:ÃÄ� ¤ÅÆÇÈ5Ci:ÃÄ� ¤�N,-:;

3.¾¿��S0µ89­®$¼+5É�Êˤ¾¿[CHIÃÄ�¤ÅÆÇÈ$µ¶;É�Êˤ¾¿+=[­®K5ÌÍA9Î(,�N$¾¿&:Ï�,-:;É�Êˤ¾¿$n)9ÐÑ=[ÒÓ0ÔÕ9­®

Page 6: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

����������� �����������������K$�NJ¾¿5� A!&)9�,-:;�9[É�Êˤ¾¿$n):*+5H.[­®KJ6)�N0¾!�:"#$$%&*+0,@[�N&J'($$)*A[+,-4·�.5/:J$01*+0,@:;ÃÄ�¤ÅÆÇÈ+=[­®KJ1)�N$�J��2345À¨©5%&Ï�,-:;*JÏ�5H8g[­®KJ1)�N0XY-5�5H8g67&:*+$01*+0,@:;4. 89§ef\],=:;<^K0m^K,-:9�[m^K

GA5¹lA989=>Ê�?$n):@A0-:;�*,[§ef\],=©%-4Ï�,-:BÊUC89 (BLX-α) 11)

$n)9;BLX-α=[2QJD�NJEdi

$αFjG5H!A9FG5[Z�N$��&:Ï�,-:;BLX-α

$n):*+5H.[D�N5I9Z�N$��,@:;BLX-J5H:MDJ89JÍ$K 75[/'CHI<'J89JÍ$K 8

5�&;Parent2

Parent1

dChild1

Child2

Fig. 7. Crossover for hue.

0.0 1.0

Parent1 Parent2

d

Child1 Child2

Fig. 8. Crossover for saturation and brightness.

5. LM<N��O�J:;<^5C)g[LM<NP5��@QUR]5:;<^K$<l#S:;§ef\],=[:;<^J^$NV

+&:+[TgJ<^5¯

A 1NV

JUP,LM<N$µ ¶;LM<N5H.[D�NJXYZVi,=��#�4)Z�N0��#�:;6. WXYZWX¨©=��SJ[\+&:;��S0]�:RSTU0c�#�9+Y^A9_²,.`a$WX&:;5. bcdefghijklmnopwqrswtu5.1 vwxyM/z{|5��?���N����0[IGA

5Ci:��SJG}~�5��,-:"$��&:9�m�$µ89;m�,=[���N����JN4:2QJef\]$n)gÌ�$µ89;4C[§m�5Ci:���=20�©J�� 20�,-.[ef\]J�n��=���Ó,N4:��+A9;m�,= ��)g�9)=� fJ�M$RSTU&:�+)¶¦U��¤$���5YÕ[Wef\]$n)gRSTU$µ8g¼E89;�9[m�WX�5=�U��¤$m�A9;m�,n)9ef\]=��J 2 Q,-:;2 QJef\]J�)=���N����J�,-.[�J�J¾¿[89[CHILM<NJXY-5���Q��?=� ,-:;ef\]J�Q��?$Table 1

5�&;• ¡Q�¢�£¤�ef\]M/z{|5��)g���N��$µ ¶ef\],-:;��S=[��J�@4M$1M�Ñ5��&:;�J�[ef\]0��S5�E�9M5�M��$�nAg��A9 3M�M$n)g 9�J���N$��&:;•

QUR]ef\]�¥-5 IGA,=[QUR]5���N$��&:;�J9�[§m�,=Ì�ef\]+Ag���N$QUR]5��&:ef\]$n):;9VA[9�N+)¶¦4)���N&J§"E��SJ¨�5(¶©)�N$��&:J=ª«,4)+¬Õ[T�N ­9�N®$̄ ^°��l,@:±#

Page 7: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

���Z ��� �F ��� ��� �$:i[��SJ¨�5�(&:���N&"ERSTU$��,@:H¶5A9;Table 1. Parameter.

Number of individuals 9

Number of design variables 9

Number of search generations Arbitrary

Crossover rate NP−NE

NP

Mutation rate 1NV

NP :Number of individuals

NE :Number of elite individuals

NV :Number of design variables§m�,�U��¤5H8g��&:Á�=��J�.,-:;��1ef\]$n)g��A9RSTU53J?E)��Ag):"��

2 ����5C)g��JRSTU='?c�Ag)9"��3ef\]$n)gRSTU$µ¶J=3J?E)�A"89"�U��¤Á�1J°�Á�= ����[������[�3¸E,¼4)�[������[����J 5³´,-.[�U�� ¤Á�2J°�Á�= �c�Ag)9�[�3¸E"+)Õ!c�Ag)9�[�3¸E,¼4)�[�3¸E"+)Õ!c�Ag)4"89�[�c�Ag)4"89�J 5³´,-:;�9[�U��¤Á�3J°�Á�=��A"89�[�3¸E"+)Õ!�A"89�[�3¸E,¼4)�[�3¸E"+)Õ!�A?4"89�[��A?4"89�J 5³´,-:;*�EJ�U�� ¤Á�J��$¼+5[M/z{|$ IGA J���N��5�n&:*+0[��SJ¨�5(¶���NJ��5��,-:"[�9[��SJG}~�5��,-:"5Q)g��$µ ¶;

5.2 vw��� !"#�U��¤Á�1"E

3J��$��O�Fig. 9"E

Fig. 115�&;�9[�U��¤Á�1

5C)g= ��

��¼A?= ������+°�A9���^+ ����¼A?= ������+°�A9���^[�U��¤Á�25C)g= �c�#�g)9�¼A?= ���c�#�g)9�+°�A9���^+ �c�#�g)4"89�¼A?= ���c�#�g)9�+°�A9���^[�U��¤Á�3

5C)g= ��A"89�¼A?= ����A"89�+°�A9���^+ ��A?4"89�¼A?= ����A?4"89�+°�A9���^$n)g[�\$% 3%,&'�Z$µ89;&'�ZJ��$Table 2

5�&;Satisfactory

Satisfactory so-so

Can't really say

Dissatisfied

Rather dissatisfied

Random system

30%

6 people

Color harmony system

65%

13 people

15%

3 people

80%

16 people

5%1 people

5%1 people

Fig. 9. Result of satisfactory level for the proposed

system.

Yes

Somewhat yes

Can't really say

No

Somewhat no

Random systemColor harmony system

15%

3(15%3 people

40%

8 people15%3 people

30%

6 people

10%2 people

15%

3 people30%

6 people

35%

7 people

Fig. 10. Result on the indicated of many preferred in-

dividuals in the initial generation.

Fig. 9 J�U�� ¤Á�15C)g[)ef\],

95% ­19�®J���0 �����9= ������+°�A9;&'�Z$µ89��[)ef\]5C)g[����¼A?= ������+°�A9���^+ ������¼A?= ����+°�A9���^JÓ5=�\4E0-89;*J*+"E[���N��J��5""BE���S=��'J1)RSTU$��,@9+)Õ:;

Page 8: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

����������� �����������������Delightful

Delightful so-so

Can't really say

Dreary

Rather dreary

Random systemColor harmony system

20%

4 people

60%

12 people

15%

3 people

5%1 people

35%

7 people

25%

5 people10%

2 people

25%

5 people

5%1 people

Fig. 11. Result on the pleasantness of the design pro-

cess in the proposed system.

Table 2. Result of sign test.

Evaluation item Significance

probability

Item 1 Color harmony system 1.90 × 10−5

Random system 9.53 × 10−7

Item 2 Color harmony system 9.44 × 10−2

Item 3 Color harmony system 1.84 × 10−3

Random system 7.08 × 10−2

Fig. 10 J�U�� ¤Á�25C)g[QUR]ef\]5¯A[¡Q�¢�£¤ef\]J�0���N5��SJ¨�5(¶�N0'?c�#�g)9*+0B":;A"A[&'�Z$µ89��¡Q�¢�£¤�ef\]5C)g �c�#�g)9�¼A?= �3¸E"+)Õ!c�#�g)9�+°�A9���^+ �3¸E"+)Õ!c�#�g)4"89�¼A?= �c�#�g)4"89�+°�A9���^JÓ5=�\4E0�E�4"89;*JA�+Ag[m�,���5¯Ag ��@4M$�M��Ag?V#)�+YÕgM$��Ag¼E89*+0¬ÕE�:;��, ��@4M�+)8g¼�JZ�=-)�),-.[��5n)9)�@4M+���Ó5n)9)�@4M0N4:43[� Ò,-8g¼¯�+&:¼J5H.�@4M�n)9)M=N4:+¬ÕE�:;¡Q�¢�£¤�ef\],=[n)g):�M��J(ª[��[¼A?=IJM/�¤�U$��(BS9�M01UP,��#�:;&4B¸[��S0ef\],·�5��A9

��J�@4M5¯Ag[�JM+MD[CHI¤�UJ�)M$'nA9RSTU0���N+Ag'?c�#�:;�J9�[�@4M+Ag��A9M+[§m�5Ci:RSTUJ¦U�� ¤,-:���Ó5n)9)+¬Õ:M0N4:��S5+8g=[¨�5(¶���N0-�.c�#�4"89J,=4)"+¬ÕE�:;�ªJ*+"E[cdÏ�,= ��@4M�,4? �RSTUJ¯�+4:¼J5n)9)�@4M�$·�5��S5��Ag¼E¶*+5H.[���N5��SJ¨�5(¶�N$'?c�&:*+5C)g#E5�n54:+¬ÕE�:;Fig. 11 J�U�� ¤Á�3

5C)g[¡Q�¢�£¤�ef\],= 80% ­16�®[QUR]ef\],=60% ­12�®J���0 ��A"89�¼A?= �3¸E"+)Õ!�A"89�+°�A9;&'�Z$µ89��[¡Q�¢�£¤�ef\],=�\E0�E�90[QUR]ef\],=�\E0�E�4"89;�9[TU? ��z�H. �¡Q�¢�£¤�ef\],=[M$��&:J0�A"89�43J\�$���"E�9;*�EJ*+"E[¡Q�¢�£¤�ef\],=���=�A�,RSTU$µÕg):*+0B":;§m�5Ci:W���J`a¨©^$ Fig. 12

5�&;

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

User

2 3 5 6 9871 4 13121110 161514 1917 18 20

������ ����������� ������� ������

Num

ber

of genera

tions

Fig. 12. Number of search generations.T���J !`a¨©^=[¡Q�¢�£¤�ef

Page 9: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

���Z ��� �F ��� ��� �\],=9.15

¨©[QUR]ef\],=12.1¨©,-.[¡Q�¢�£¤�ef\]J�0¦4?489;�9[�J�%�E=¡Q�¢�£¤�ef\],=4.46[QUR]ef\],=

5.63,-.[¡Q�¢�£¤�ef\]J�0�#?489;*�EJ*+5�Õ[Fig.

12$�g¼B":+C.[¡Q�¢�£¤�ef\]=QUR]ef\]5Ì�[Ì�-�ZA9`a¨©^,RSTU$��,@:+¬ÕE�:;�ªJ*+"E[IGAJ���N��$M/z{|5��@µ¶¡Q�¢�£¤�ef\],=[��SJ¨�$01)RSTU$µÕ[#E5RSTU$��S0�A�,µ¶*+0,@g):*+"E[��SJ�ÐG}$~�,@9+)Õ:;�9[QUR]ef\]+Ì�Ag[¦4)`a¨©^,RSTU$µÕg):*+"E[���N��5M/z{|$�n&:*+5H.[.`aJ�Pl0K�[��SJG}J~�5��,-89+)Õ:;

6. ��y§��,=[IGAJ.`a5Ci:��SJG}~�$�-+A[��SJ$$� #S[¨�5( ¶���N$��&:9�JÏ�$cdA9;��SJ¬5( ¶���N$��&:*+5H.[­®J�5��S5ÔÕ:�Ð-G}${E�["Q��'J1).$��&:*+$K89; N-5=[�JÏ�+Ag�Mz{|5��?���N����$cdA9;cdÏ�,=[��S0�Ñ5��A9�M$¼+5�M��$�nA[��#�9�M$n)g���N$��&:;�9[cdÏ�J��$$��&:9�[�ÒJ���ÓJ�MRSTU$µ ¶=� f�ÓRSTUef\]$O�A9;��m�J��[cdÏ�,=��S=��'J1)RSTU$��,@:*+0B"89;�9[QUR]5���N$��&:�(+Ì�Ag[¦4)`a¨©^,RSTU$��,@[�Q�A�,RSTU$µ¶*+0,@9*+�E[cdÏ�= IGA5Ci:��SJG}~�5��,-:*+0B�89;

� " � �1) �§±���. HCD ¢UCBË¢ ÒÓ§�:;.

2006.

2)����

. �l;�5��?��-:;|. � !��", Vol. 64, No. 10, pp. 1419–1422, 1998.

3)1�#µ

, $�%&, '()*.¯+,�l;��J��-..

Ò!#��", Vol. 13, No. 5, pp.

692–703, 1998.

4) D.E.Goldberg. Genetic Algorithms in Search Op-

timization and Machine Learnig.

5)1�#µ

, $�%&, '/)*.TU?Q¢\ B�l;�[XY-�É0Ä1]4. 23K4, pp. 325–361,

2000.

6) Hideyuki Takagi. Interactive Evolutionary Compu-

tation: Fusion of the Capabilities of EC Optimiza-

tion and Human Evaluation. Proceedings of IEEE,

2001.

7) 5��,1�#µ

.¯+,

ga5H:

3À6

cgQT\ U7RSTU89

. :Z_`����|;",

Vol. J81-D-2, No. 7, pp. 1601–1608, 1998.

8)���F

, <�=Z, >?@AZ,{BCD

, E�FG,��

.¯+,XY-�É0Ä1]$n)9HTUIhJ��ef\]. K��LUTU?�M�f��|;", Vol. 10, No. 2, pp. 243–251, 2008.

9) N§O&. M/�J�P.;l4Q=;R

, pp. 227–

244, 1997.

10) S T� ­U®,VW�Ò�§M/��, ­X®.RY?ÉM/L¤��É. 2004.

11) L.J Eshleman and J.D Schaffer. Real-Coded Ge-

netic Algorithms and Interval-Schemata. Founda-

tions of Genetic Algorithms, Vol. 2, pp. 187–202,

1993.

Page 10: Interactive Genetic Algorithm using Initial …mikilab.doshisha.ac.jp/achieve/data/2011/resume/rikoken...Interactive Genetic Algorithm using Initial Individuals Generated from Human

����������� ��������������������� ����p

Pale

ltgLight

Grayish

g Grayish

vVivid

bBright

sStrong

dpDeep

ltLight

sfSoft

dk Dark

dDull

dkgDark

Grayish

WWhite

ltGyLight Gray

mGyMedium Gray

dkGyDark Gray

BkBlack

Saturation HighLow

Brightn

ess

Low

High

2:R

3:yR

4:rO

5:O

6:yO

7:rY8:Y

9:gY

10:YG

11:yG

13:bG

14:BG

12:G

15:BG

16:gB

17:B

21:bP

24:RP

23:rP

22:P

20:V19:pB

18:B

1:pR

��� �� �����yellowish red

����� �reddish orange

� ��� �� � !!"# $% &!'' ()red

yello

wish

gree

n

green

blu

ish p

urp

le

redd

ish

purp

lepurplish red

yello

w g

reen

red purple

purp

lish b

lue

blue green

blue

bluish green

purp

le

vio

let

*+,,-./0 +1123greenish blue

Parent2

Parent1

dChild1

Child2

Fundamental color

Faux-CamaieuTone on Tone

Gradation Natural harmony

Fig. 3. Tone of PCCS. Fig. 4. Hue circle of PCCS.

Fig. 5. Example of color combination.

Fig. 6. Method of generating initial individuals

based on color the harmony theory.

Fig. 7. Crossover for hue.