Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
NSGA
Search
kichinosukey
December 12, 2018
Science
0
55
NSGA
参考:
https://web.njit.edu/~horacio/Math451H/download/SrinivasDeb_GA.pdf
kichinosukey
December 12, 2018
Tweet
Share
More Decks by kichinosukey
See All by kichinosukey
カーネル関数とは
kichinosukey
0
54
Latin Hypercube Sampling
kichinosukey
0
73
Lasso.pdf
kichinosukey
0
56
NSGAII
kichinosukey
0
50
基底関数回帰の概要について
kichinosukey
0
77
児童虐待の現状
kichinosukey
0
61
イノベーションのジレンマとフィンテック
kichinosukey
0
73
Paxos made simpleを理解しよう
kichinosukey
0
510
Other Decks in Science
See All in Science
Lean4による汎化誤差評価の形式化
milano0017
1
410
データから見る勝敗の法則 / The principle of victory discovered by science (open lecture in NSSU)
konakalab
1
260
Performance Evaluation and Ranking of Drivers in Multiple Motorsports Using Massey’s Method
konakalab
0
130
HDC tutorial
michielstock
1
320
Text-to-SQLの既存の評価指標を問い直す
gotalab555
1
160
MCMCのR-hatは分散分析である
moricup
0
570
(2025) Balade en cyclotomie
mansuy
0
400
Collective Predictive Coding as a Unified Theory for the Socio-Cognitive Human Minds
tanichu
0
150
【RSJ2025】PAMIQ Core: リアルタイム継続学習のための⾮同期推論・学習フレームワーク
gesonanko
0
580
Accelerating operator Sinkhorn iteration with overrelaxation
tasusu
0
160
Rashomon at the Sound: Reconstructing all possible paleoearthquake histories in the Puget Lowland through topological search
cossatot
0
400
機械学習 - SVM
trycycle
PRO
1
960
Featured
See All Featured
Breaking role norms: Why Content Design is so much more than writing copy - Taylor Woolridge
uxyall
0
130
Building Applications with DynamoDB
mza
96
6.9k
From Legacy to Launchpad: Building Startup-Ready Communities
dugsong
0
130
Agile Actions for Facilitating Distributed Teams - ADO2019
mkilby
0
100
Conquering PDFs: document understanding beyond plain text
inesmontani
PRO
4
2.2k
Claude Code のすすめ
schroneko
67
210k
DevOps and Value Stream Thinking: Enabling flow, efficiency and business value
helenjbeal
1
83
Done Done
chrislema
186
16k
Designing for Timeless Needs
cassininazir
0
120
The innovator’s Mindset - Leading Through an Era of Exponential Change - McGill University 2025
jdejongh
PRO
1
82
We Have a Design System, Now What?
morganepeng
54
8k
Side Projects
sachag
455
43k
Transcript
/4("
ଟత࠷దԽ ↟Y1ᮈᕓຌแᨷม ↟OPOEPNJOBOU TVQFSJPSᓓᙚֶ᧣ͼḑᶡḿ Min/Max fi (x) … i =
1,2, … , N gi (x) … j = 1,2, … , N hk (x) … k = 1,2, … , N
࠷খԽʹ͍ͭͯ ↟ͥᙉfYYΤᬧ፺ EPNJOBUFT ͖ͩΝ ↟ΐͱYYͽରͩᅿ͵͖ Ν CFJOGFSJPSUP ͺ͖͘ ↟ᝂͽ᷁ͽΔᬧ፺ͧΞ͖ͼ͖f Δͩ͡ᅿ͵͖ͼ͖ᦉᷯ
OPOEPNJOBOUOPOJOGFSJPSͺ ͖͘ ↟ᱠ᧱ặዧ᩼ḑOPOEPNJOBOU ͼḑͥͺΤ͖͘ Y Y YJTTUSJDUSZHSFBUFSUIBOY
ݹయతͳղ๏ ↟᧣ᬍΑͣ ↟ᬍΑΤᑿᠭ ↟ฌሜؔ ↟ጽ᧣ͽϫ⑲δϯψώฌሜΤᠭ͖Ν ↟ᱠႝfᱠᖀ໋ᙚ ↟ᄑ͖ͩᠸᔬႷᶝᱠ᧱ặ
తͷॏΈ͚ ↟᧣ᬍ᠓ᕥͽԠͪᬍΑϞδύϰΤᕅ ↟ᓓᙚΚΜfᬍ᠓ႷΤວᅝͩͱϖϱ⑲ύḑΤᄌΝͥͺ͟͞Ν Z = N ∑ i=1 wi fi
(x) where x ∈ X 0 ≤ wi ≤ 1, N ∑ i=1 wi = 1
ڑؔ ↟᧣ؔᤎ᷀ͺͼΝfᴢᓝSďϫ⑲δϯψώฌሜĐ ↟ZᶝͽରͫΝ෪ମᱠີᲩ Z = [ N ∑ i=1 |
fi (x) − yi |r ] 1/r , 1 ≤ r ≤ ∞
NJONBY ↟ᄑ͖ͩᠸᔬႷᶝΤᱠ᧱ặ minimize Fi (x) = max[Zi (x)] Zi (x)
= fi − fi fi , i = 1,2, … , N
͜ΕΒͷܽ ↟͖ͬΞΔ୯᧣Τᓡᨷͩ͜Μḿ࣮ͺ༏ሜ ↟ሜᒺมͽରԠ͟ͼ͖ ↟ᬍΑͺ᠓ཏϱϞϰďแᨷมĐͽରͩᑤᓝͽπϸμϋΨϛ ↟ͥΞ͞Ჺ᧣fჯͪዧ᩼Τ᷁ႷΔΜጿͩḑ͡ᡊͽͼΝ
୯७త ↟ͥᦉᷯϖϱ⑲ύḑͻͥͽ͔Ν ͝Ħ ↟Y᎘ғͽ͔Ν ↟ᬍΑΤͯΞͰΞ᧣ͽରͩ X
ͺม͚Ν ↟ͫΝͺϖϱ⑲ύḑΔY Y Yͺ Y᎘ғḑ͖͞͝Κ͘ͽΔ ხ͖ͩΐ͘ ↟ͶΐΜfᑤแᨷ᧣͔ΝͺΔᜡ͚Ν Minimize f11 = x2 Minimize f12 = (x − 2)2
(FOFUJD"MHPSJUIN ↟3PTFOCFSHၞ᧣Τጩ͘୯ᕺᶈᓺዲ᭽ᷯ͞ပͥͫặֶ มặᄑμϤϪϱ⑲μϬϸᓷͽରͫΝᩮᚳ ↟4DIBGGFS࣮ᠭ᧣ͼΧϰιϯοϥ 7FDUPS&WBMVBUFE(FOFUJD "MHPSJUIN ↟(PMECFSH7&("ᛰᨨΤᏎ͘ͱΓOPOEPNJOBOUTPSUJOH Ϝϲμ⑲νϨΤᩮᚳ ↟/4("ͥΞͽͷ͖͖Ν
/4(" ↟᳔ᩮᚳͺfᕅίϟϱ⑲τ⑲ხᥭͽ͖͜ΑᤘͼΝ ↟༜ͧfมᤘ᳔("ͺჯͪ ↟ᗌᘡ ↟'JSTU3BOLΤᨷᤅf᧱ᷯႷυϤ⑲ᲩΤ'᭽ஂͽ༩͚ΝfͯΞͰΞͽ ରͩᅿᲩGJBΤຌᒿf'᭽ஂΤ͖᩹ͱΜ෪ମͽରͩჯ༷ ॲሑΤΜጿͫfGGjGJ
/4(" ↟ᔬͻॲሑΤᕴႚΜጿͩḝ͘ͽ͔ͱ͵f᭽ஂᦪᓺᒽ ͧΞΝ ↟උϚϲϸύΤͻᷡᷯᦪᓺᒽͫΝ͝᧱ᷯႷͽै͘ ↟'Ϯϸδ෪ମ᳔ΔΚΜΔ᧱ᷯႷ͞ີ͖ͱΓၞ͡θϙ⑲ΤᥭΝ ͥͺ͟͞Ν ↟ͥΞͽΚΜϖϱ⑲ύḑΊऩᖣΤ᱉ͫͥͺ͞൱ၙ ↟ϚϲϸύϮϸδΤ๕ͽ᭽ஂ۠Ꭽ͞൱ၙ
/4("
/4("ͷϑϩʔνϟʔτ ↟μάΧϯϸε ↟ၞ༷ᕥệᏂͱΓᬥᶝ ↟෪ମJ KණฌሜͺȀϖϮϦ⑲τͽΚ͵ᶡḿ ͧΞΝ ↟ͻͳΛΔQIFOPUZQJD ↟ͯΔͯΔͥΚ͘ͼ᪣ᨿ൱ၙͼϖϮϦ⑲τ Τᐾ͚ͩΐ͵͖Νͥͺዧ᩼