ๆญฏ98419515.PDF



Similar documents
๋‚จ๋ถํ•œ๊ต๊ณผ์„œ์—์„œ๋‚˜ํƒ€๋‚œ ๋ฏผ์กฑ์ •์ฒด์„ฑ

ๆญฏFFF01379.PDF

ๆญฏ320.PDF

์ด์ˆ˜์ง„.PDF

์ •๋ด‰์ˆ˜.PDF

ยณยปรรถ

์ˆ˜ํƒ์—ฐ๊ตฌ01-09(์ˆ˜์š”์ž ์ค‘์‹ฌ1).hwp

ๆญฏ๊ฐ•์œค์‹.PDF

ๆญฏ PDF

<C1DFB1DE2842C7FC292E687770>

์ธ ์‚ฌ ์ฒญ ๋ฌธ ์š” ์ฒญ ์‚ฌ ์œ  ์„œ

<3130BAB9BDC428BCF6C1A4292E687770>

ๆญฏ์—ฐ๋ณด01-08.PDF

ๆญฏ7๊ถŒ2ํ˜ธ.PDF

ๆญฏ PDF

ePapyrus PDF Document

ๆญฏJ01-13.PDF

S hip Plannin g S y s tem for Efficient Cran e Operation of Cont ain er T ermin al

a16.PDF

105รˆยฃ4fลก

ๆญฏD13236_F1.PDF

ๆญฏ์—ฐ๋ณด01-04.PDF

2/21

untitled

๋‚ด์ง€4์›”์ตœ์ข…

6 ๊ฐ•๋‚จ๊ตฌ ์ฒญ๋‹ด์ง€๊ตฌ ์ฒญ๋‹ด๋™ 46, ์‚ผ์„ฑ๋™ 52 ์ผ๋Œ€ 46,592-46,592 7 ๊ฐ•๋‚จ๊ตฌ ๋Œ€์น˜์ง€๊ตฌ ๋Œ€์น˜๋™ 922๋ฒˆ์ง€ ์ผ๋Œ€ 58,440-58,440 8 ๊ฐ•๋‚จ๊ตฌ ๊ฐœํฌ์ง€๊ตฌ ๊ฐœํฌ๋™ 157์ผ๋Œ€ 20,070-20,070 9 ๊ฐ•๋‚จ๊ตฌ ๊ฐœํฌ์ง€๊ตฌ์ค‘์‹ฌ ํฌ์ด๋™ 238 ์ผ๋Œ€ 25,070-25,

27์ง‘์ตœ์ข…10.22

ํ™ฉ๋ฃก์‚ฌ ๋ณต์› ๊ธฐ๋ณธ๊ณ„ํš โ…ฅ. ์‚ฌ์—ญ ๋ฐ ์ฃผ๋ณ€ ์ •๋น„๊ณ„ํš ๊ฐ€. ์‚ฌ์—ญ์ฃผ๋ณ€ ์ •๋น„๊ตฌ์ƒ ๋ฌธํ™”์œ ์ ์ง€๊ตฌ ์กฐ์„ฑ 1. ์ •๋น„๋ฐฉํ–ฅ์˜ ์„ค์ • ํ™ฉ๋ฃก์‚ฌ ๋ณต์›๊ณผ ํ•จ๊ป˜ ์ฃผ๋ณ€ ์ž„ํ•ด์ „์ง€(์•ˆ์••์ง€) ๆตทๆฎฟๅ€(้›้ดจๆฑ )์™€ ๋ถ„ํ™ฉ์‚ฌ ๋“ฑ์˜ ๋ฌธํ™”์œ ์ ๊ณผ ๋„คํŠธ์›Œํฌ๋กœ ์—ฐ๊ณ„๋˜๋Š” ์ข…ํ•ฉ์  ์ •๋น„๊ณ„ํš์•ˆ์„ ์ˆ˜๋ฆฝํ•œ๋‹ค. ์ฃผ์ฐจ์žฅ๊ณผ ๊ด‘์žฅ ๋“ฑ ์ฃผ๋ณ€

ๆญฏ95_dbtech.PDF

1์ฐจ๋‚ด์ง€

ๆญฏ (์ด์žฌ์Šน)[1].PDF

1-1ร‡ร˜ยผยณ

<BACFC7D1B3F3BEF7B5BFC7E22D3133B1C733C8A BFEB2E687770>



<BCD2B0E6C0FCBCAD20C1A62032B1C72E687770>

์ž…์žฅ

<C7D1B1B9B0E6C1A6BFACB1B8C7D0C8B828C0CCC1BEBFF85FC0CCBBF3B5B75FBDC5B1E2B9E9292E687770>

UXJQXAAOAIYR.hwp

DBPIA-NURIMEDIA

ๆญฏ PDF

ๆญฏIC-706.PDF

yesแ„‰แ…ตแ„‹แ…กแ†ซ1007_แ„Žแ…ฌแ„Œแ…ฉแ†ผ_2_30

( )๋ฐ•์šฉ์ฃผ97.PDF

01. H H ํ˜•๊ฐ• s and al Properties ์น˜์ˆ˜๋ฐ๋‹จ๋ฉด์„ฑ๋Šฅ (1) Metric Series - KS, JIS '90 Division (depth x width) x x x x x 300

ALPHA I 500M(ํ˜„๋Œ€์‚ผํ˜ธ) service manual(kor).hwp

<C0CEB9AEBBE7C8B C1FD32C8A3292DC3D6C1BE2E687770>


์„ธ๊ณ„ ๋น„์ง€๋‹ˆ์Šค ์ •๋ณด

ๆญฏJ02-10.PDF

PDF

<4D F736F F D20C3A520BCD2B0B32DB4D9B0A1BFC0B4C220B0E6C1A6C1F6C1F8322E646F63>

No Title

(71) ์ถœ์›์ธ ๋‚˜ํ˜œ์› ๋Œ€๊ตฌ ๋‹ฌ์„œ๊ตฌ ๋„์›๋™ 1438 ๋Œ€๊ณก์‚ฌ๊ณ„์ ˆํƒ€์šด ๋‚˜ํ˜œ๋ฆฌ ๋Œ€๊ตฌ ๋‹ฌ์„œ๊ตฌ ๋„์›๋™ 1438 ๋Œ€๊ณก์‚ฌ๊ณ„์ ˆํƒ€์šด (72) ๋ฐœ๋ช…์ž ๋‚˜ํ˜œ์› ๋Œ€๊ตฌ ๋‹ฌ์„œ๊ตฌ ๋„์›๋™ 1438 ๋Œ€๊ณก์‚ฌ๊ณ„์ ˆํƒ€์šด ๋‚˜ํ˜œ๋ฆฌ ๋Œ€๊ตฌ ๋‹ฌ์„œ๊ตฌ ๋„์›๋™ 1438 ๋Œ€

ๆญฏFFF01288.PDF

<33312D312D313220C0CCC7D1C1F820BFB0C3A2BCB12E687770>

ๆญฏ6๊ถŒ2ํ˜ธ.PDF

h99-37.PDF

<313120C0AFC0FCC0DA5FBECBB0EDB8AEC1F2C0BB5FC0CCBFEBC7D15FB1E8C0BAC5C25FBCF6C1A42E687770>

untitled

๋…ผ๋ฌธ์ˆ˜์ •๋ณธ.PDF

untitled

<C5F0B0E82D313132C8A328C0DBBEF7BFEB292E687770>

hwp

๋ถ€๋ฌธ๋ณ„ ์—๋„ˆ์ง€์› ์ˆ˜์š”์˜ ๋ณ€๋™ํŠน์„ฑ ๋ฐ ๊ณตํ†ต๋ณ€๋™์— ๋ฏธ์น˜๋Š” ๊ฑฐ์‹œ์  ์š”์ธ๋“ค์˜ ์˜ํ–ฅ๋ ฅ ๋ถ„์„

main.hwp

< C7D0B3E2B5B520B9FDC7D0C0FBBCBABDC3C7E820C3DFB8AEB3EDC1F528C8A6BCF6C7FC292E687770>

<C1A4C3A5BFACB1B D3420C1A4BDC5C1FAC8AFC0DAC0C720C6EDB0DFC7D8BCD220B9D720C0CEBDC4B0B3BCB1C0BB20C0A7C7D120B4EBBBF3BAB020C0CEB1C720B1B3C0B020C7C1B7CEB1D7B7A520B0B3B9DF20BAB8B0EDBCAD28C7A5C1F6C0AF292E687770>

ๆญฏjeong-shik.PDF

<C5D8BDBAC6AEBEF0BEEEC7D C1FD2E687770>

DBPIA-NURIMEDIA

ๆญฏ๋ชฉ์ฐจ13.PDF

์ตœ์ข…๋ณด๊ณ ์„œ.PDF

๋ณด๊ด€๋ณธ.PDF

334 ้€€ ๆบช ๅญธ ๊ณผ ๅ„’ ๆ•Ž ๆ–‡ ๅŒ– ็ฌฌ 55 ่™Ÿ ่ง’ ่ชช ์—์„œ๋Š” ๋ฟ”์ด ๋‚œ ๋ง๊ณผ ๊ณ ์–‘์ด๋ผ๋Š” ๊ธฐํ˜•์˜ ๋™๋ฌผ์„ ์†Œ์žฌ๋กœ ํ•˜์—ฌ ๋‹น๋Œ€ ์ •์น˜ ์ƒ ํ™ฉ์„ ๋น„ํŒํ•˜์˜€๊ณ , ็™ฝ ้ป‘ ้›ฃ ์—์„œ๋Š” ์„ ๊ณผ ์•…์„ ์ƒ์ง•ํ•˜๋Š” ์ƒ‰๊น”์ธ ็™ฝ ๊ณผ ้ป‘ ์ด ์„œ๋กœ ๋ฒŒ์ด ๋Š” ๋ฌธ๋‹ต์„ ํ†ตํ•˜์—ฌ ์˜ณ๊ณ  ๊ทธ๋ฆ„์˜ ๊ฐ€์น˜๊ด€์ด ์ „๋„๋œ ํ˜„์‹ค์„ธ

PDF

<C7D0B1B3C7F5BDC520BBE7B7CAB9DFB1BCB0FA20C8AEBBEAC0BB20C0A7C7D120B3D7C6AEBFF720B1B8C3E0B9E6BEC8BFACB1B D30362C20C0CEBCE2BABB292E687770>

ๆญฏ PDF

LG1-P0673.eps

???? 1

ๆญฏPLSQL10.PDF

ยฑยนยนรŽร€ยบร‡ร ร‡รถรˆยฒ.PDF


ๆญฏ PDF

09๊ถŒ์˜ค์„ค_ok.hwp

< BACFC7D1B1B3C0B0C1A4C3A5B5BFC7E228B1E2BCFABAB8B0ED D D20C6EDC1FD2035B1B32E687770>


1. ํ™”์„ฌ์‚ฐ์—…์˜ ๊ฐœ์š” 1.1 ํ™”์„ฌ์‚ฐ์—…์˜ ๋ฅต์„ฑ ํ™” ์„ฌ์‚ฐ ์—… ์˜ ์‚ฐ ์—… ์  ํŠน ์„ฑ ํ™”์„ฌ์‚ฐ์—…์€ ์›์‚ฌ,์›๋ฉด ๋‘ฅ ๊ธฐ์ดˆ์†Œ์žฌ๋ฅผ ์ƒ์‚ฐํ•˜๋Š” ์„ฌ์œ ์‚ฐ ์—…์˜ ํ•ต์‹ฌ์‚ฐ์—…์œผ๋กœ ๊ณ ์šฉ์ฐฝ์ถœํšจ๊ณผ์™€ ๊ณ ๋ถ€๊ฐ€๊ฐ€์น˜๋ฅผ ์‹คํ˜„ํ•  ์ˆ˜ ์žˆ์Œ O ์„ธ๊ณ„์ ์œผ๋กœ ๋…์ผ,์ดํƒœ๋ฆฌ,์ผ๋ณธ ๋“ฑ ์„ ์ง„๊ตญ์„ ์ค‘์‹ฌ์œผ๋กœ M E(๋งˆ์ดํฌ๋กœ ์ผ๋ ‰ํŠธ


CONTENTS.HWP

INDUS-8.HWP


ยผยบยฟรธรรธ รƒรขยทร‚-1

์ดˆ๋ณด์ž๋ฅผ ์œ„ํ•œ ์ž๋ฐ” 2 21์ผ ์™„์„ฑ - ์ตœ์‹ ๊ฐœ์ •ํŒ

[ReadyToCameral]RUFยนรถร†ร›(CSTA02-29).hwp

( )์‹คํ—˜๊ณ„ํš๋ฒ•-๋จธ๋ฆฌ๋ง ok

ํƒ„๋„๋ฏธ์‚ฌ์ผ ๋ฐฉ์–ด๋ฌด๊ธฐ์ฒด๊ณ„ ๋ฐฐ์น˜๋ชจํ˜• ์—ฐ๊ตฌ (Optimal Allocation Model for Ballistic Missile Defense System by Simulated Annealing Algorithm)

NX Nastran brochure (Korean)

์†ก๋™์šฐ.PDF

Transcription:

A Study on E ffect s of Selection Schem es in Genetic Progr am ming for T im e Series Prediction 2000 2

A Study on E ffect s of Selection Schem es in Genetic Progr am ming for T im e Series Prediction 1999 10 1999 12

1. 1 1.1 1 1.2 3 2. 4 2.1 4 2.2 5 2.3 7 3. 10 3.1 10 3.2 12 4. 16 4.1 16 4.2 18 5. 22 5.1 22

6. 27 28

1. 1.1 (selection ).,. (selective pressure) [Bäck - 96].. (genetic diversity ),... (evolution strategy ) (ranking selection ). (tournament selection ) (evolutionary programming ). (genetic algorithm ) (proportional selection). Goldberg Deb (convergence time) (growth ratio) [Goldberg - Deb - 91]. Back Hoffmeister (takeover time) (genetic diver sity ) [Bäck- Hoffmeister - 91]. Mühlenbein Schlierkamp- Voosen (selection differential) (response to selection) [Mühlenbein- SV - 93].. Blickle T hiele - 1 -

(selection variance) [Blickle- T hiele- 95]. (genetic programming ). - 2 -

1.2..,.. 2, 3. 4, 5. 6. - 3 -

2. 2.1.,,.,,, (, ).... - 4 -

2.2. X ( t). X ( t) = (x ( t), x ( t - 1),,x ( t - )). (1) x ( t + ) f (X ( t))., 1. = 1. = 2 = 1. t + 1 X ( t) = (x ( t),x ( t - 1), x ( t - 2)) (2) x ( t + 1). f k (normalzed mean squared error, NMSE ). - 5 -

E k = N M SE ( k) = 1 N Va r N (x ( t + 1) - f k ( X ( t)) ) 2 t = 1 N (x ( t + 1) - x)2 N t = 1 t = 1 x ( t + 1) - f k (X ( t)) 2 (3) Var = 1 N N t = 1 x ( t + 1) - 1 N N t = 1 x ( t + 1) 2 (4) NMSE (predictor ) f k (X ( t)). - 6 -

2.3. Huebner [Huebner - 89-1, Huebner - 89-2, Huebner - 89-3]. [Zhang - et - al- 97]. 81.5 14NH3 cw (FIR). aq(8,7) NH3 N2O P (13). laser 1. LeCroy.. s/ n 300. a/ d bit. (intensity pulsation )....,..,. - 7 -

1000,.,. [Zhang - et - al- 97]. 1 2. - 8 -

1 2-9 -

3. 3.1. Koza [Koza- 92].,......... 1... - 10 -

1. F T A (0). g 1. 2. D A (g ). 3. A (g ),, A ( g + 1). 4.. g g + 1. 2. 1 3-11 -

3.2. A i A. A = {A 1, A 2,,A M } (5). F = {+, -,, % } T = {x ( t),x ( t - 1),x ( t - 2),R }. % 0 1, R [ 0, 1]. 4.. f ( t) = (x ( t - 2) + 0.48) x ( t - 2) 0.48 % x ( t - 1) % (x ( t - 2) + x ( t)) x ( t) (6) - 12 -

4, A k D = { (X ( t),x ( t + 1)) } N t = 1 (7) f k (X ( t)). f k 3 NMSE E k. A k. - 13 -

F k = 1 1 + E k + (K S k ), k = 1, 2,, M (8) S k K.,......... (incremental data inheritance, IDI) [Zhang - Joung - 99- GP].. - 14 -

M G m ax R P c 200 500 20 0.9 P m 0. 1 K 0.00001 L 20 x ( t), x ( t - 1), x ( t - 2), R +, -,, % 2-15 -

4. 4.1 3,,.,. g i (selection probalilty ). p s ( A g i ) = f ( A g i ) f ( A g j ) j = 1 (9) 0... [Blickle- T hiele- 95]. q, q-.. q-. p s ( A g i ) = 1 q ( ( - i + 1) q - ( - i) q ) (10),. (, ) [Schwefel- 95] - 16 -

1,. p s ( A g i ) = { 1, 1 i 0, < i (11) - 17 -

4.2 NMSE.. 3 20 MIP 0.133 0.063 0.070 0.060 0.204 0.116 0.131 0.097 3 NMSE (MIP [Angeline- 98].) 4 5. f ( t) = ( ( ( v0 % ( ( ( ( ( ( ( v1 + v0% ( ( v2 % v0) ( v1 % ( v2 - v1))))) % ( ( ( ( v1 + ( v 1 + ( v0% ( ( v2 % v0) 0.67)))) + v1) (12) 0.81))% (v1% ( v1 0.79))) + v2) + 0.79) v2) + v0)) % ( v1% 0.51)) v2) - 18 -

5 6-19 -

7-9 NMSE.. (fitness scaling ), (tournament size) (truncation threshold).., 100 5 2 ( 9). 7 NMSE - 20 -

8 NMSE 9 NMSE - 21 -

5. [Goldberg - Deb - 91, Bäck- 96], [Mühlenbein - SV - 93], [Blickle- T hiele- 95] 3..... 5.1. Mühlenbein Schlierkamp- Voosen [Mühlenbein - SV- 93]. R (g ) g + 1 g. R (g ). R (g ) = M (g + 1) - M ( g ) (13) M ( g ) g. M s ( g ). S ( g ) = M s ( g ) - M (g ) (14). - 22 -

, S (g ) R ( g ). R (g ) = b g S (g ) (15) b g (quantitative genetics) realized heritability. b g s R s. R s = s s R ( g ) = b S ( g ) (16) g = 1 g = 1 (heritability ). 10-12 20 S (g ). 6-8 NMSE. R (g ) ( 13-15).,. - 23 -

10 11-24 -

12 13-25 -

14 15-26 -

6...,.... - 27 -

[Angeline- 98] Angeline, P. J., Ev olving predict ors for chaotic tim e series, A pp lications and S cience of Comp utational In tellig ence: P roceeding s of S PI E, Volum e 3390, S. Roget s, D. F ogel, J. Bezdek, and B. Bosacci (Eds.), SPIE, Bellingham, WA, pp. 170-180, 1998. [Bäck - 96] Bäck, T., E volutionary A lg orithm s in T heory and P ractice, Oxford Univ er sity Press, New York, 1996. [Bäck - Hoffm eister - 91] Bäck, T. and Hoffm eist er, F., Ext ended selection m echanism s in genetic algorithm s, In Belew, R.K. and Booker, L.B. (eds.), P roc. 4 th I nt. Conf. on Gen etic A lg orithm s, Morgan Kaufm ann, pp. 92-99, 1991. [Blickle- T hiele- 95] Blickle, T. and T hiele, L., A m athem atical analy sis of tournam ent selection, In P roc. S ix th In t. Conf. on Genetic A lg orithm s, Morgan Kaufm ann, pp. 9-16, 1995. [George- et - al- 94] George E.P. Box, Gwilym M. Jenkin s and Gregory C. Rein sel, T im e s eries analy s is : f orecas ting and control, Prentice Hall 1994. [Goldberg - Deb - 91] Goldberg, D. E. and Deb, D., A comparativ e analy sis of selection schem es u sed in genetic algorithm s, In Rawlin s, G.J.E. (ed.) F oundations of Genetic A lg orithm s, pp. 69-93, 1991. [Huebner - 89-1] U. Huebner, N. B. Abraham, and C. O. W eiss : ``Dim en sion s and entropies of chaotic inten sity pulsation s in a single- m ode far - infrared NH3 laser.' ' P hy s. R ev. A 40, p. 6354 (1989) - 28 -

[Huebner - 89-2] U. Huebner, W. Klische, N. B. Abraham, and C. O. W eiss : On problem s encountered w ith dim en sion calculation s. M eas ures of Comp lex ity and Chaos ; Ed. by N. B. Abraham et. al., Plenum Press, New York 1989, p. 133 [Huebner - 89-3] U. Huebner, W. Klische, N. B. Abraham, and C. O. W eiss : Comparison of Lorenz- like laser behavior with the Lorenz m odel. Coherence and Quantum Op tics VI ; Ed. by J. Eberly et. al., Plenum Press, New York 1989, p. 517 [Kim - Zhang - 98] Kim, J.- J. and Zhang, B.- T., Comparison of selection schem es for m achine lay out design, In P roc. A s ia- Pacif ic Conf. on S im ulated E v olution and L earning (SEAL- 98), Canberra, Au stralia, 1998. [Koza - 92] Koza J. R., Genetic P rog ram m ing : On the P rog ram m ing of Comp uters by M eans of N atural S election, MIT Press, 1992. [Mühlenbein - SV - 93] Mühlenbein, H. and S chlierkamp - Voosen, D., Predictiv e m odels for the breeder genetic algorithm : Continuou s par am eter optimization, E volutionary Comp utation, 1(1):25-49, 1993. [Mulloy - Riolo- Savit - 96] Mulloy, B.S., Riolo, R.L. and Savit, R.S., Dynamics of genetic programming and chaotic tim e series prediction, In Genetic P rog ram m ing 1996, pp. 166-174, 1996. [S chw efel- 95] S chw efel, H.- P., E v olution and Op tim um S eek ing, Wiley, New York, 1995. [W eigend- Gershenfeld- 94] W eigend, A. and Ger shenfeld, N., T im e S eries P rediction: F orecas ting the F uture and Unders tanding the P as t, New York : A ddison - W esley,1994 [W hitley - 89] W hitley, D., T he Genit or algorithm and selection pressure: W hy - 29 -

rank - based allocation of reproductiv e trials is best, In P roc. T hird Int. Conf. on Genetic A lg orithm s, Morgan Kaufm ann, pp. 116-121, 1989. [Zhang - Joung - 99- GP ] Zhang, B.- T. and Joung, J.- G., Genetic programming with increm ental dat a inheritance, Genetic and E volutionary Comp utation Conf erence (GECCO- 99), Orlando, Florida, Morgan Kaufm ann, 1999. [Zhang - et - al- 97] Zhang, B.T., Ohm, P., and Mühlenbein, H., Ev olutionary induction of sparse neural trees, E volutionary Comp utation, 5(2):213-236, 1997. - 30 -

.,...,.,,,,,,,., shlim,,,, ~, 94.,,,.,,.,. Hubert Wang.,,,.,. : RAT M,,,, Matt Groening,,,,,,, l' arc en ciel, rancid, bob marley, doors, X- japan,...