• No results found

Combinatorial optimization in VLSI physical design

N/A
N/A
Protected

Academic year: 2021

Share "Combinatorial optimization in VLSI physical design"

Copied!
179
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

C o m b i n a t o r i a l O p t i m i z a t i o n in V L S I

P h y s i c a l D e s i g n

bv

Pet'or A u liio n y W alsh

15.Sc. ( Ho no ur . . ) , N a t i o n a l I i v e r s i t y o f I rel and, 1982 M. Sc . ( H o n o u r s ) , N a t i o n a l H n i v e r s i l y o f I rel and, 1985 A d i s s e r t a t i o n s u b m i t t e d in parti al f ul f i l lme nt of l.lie r e q u i r e m e n t s for t h e d e g r e e of D o c t o r in P h i l o s o p h y in t he D e p a r t m e n t o f C o m p u t e r S c i e n c e W e ac c e pt this d i s s e r t a t i o n as c o n f o r m i n g t o t he required s t a n d a r d Dr. D . M. Mi l l er , S u p e r v i s o r ( D e p a r t m e n t o f C o m p u t e r S c i e n c e ) Dr. ( y T p i s , , W$ > . r t m e n t a l M e m b e r ( D e p a r t m e n t o f C o m p u t e r S c i e n c e )

—_ —^

_

---Dr. M. Se r r a, Dopnrt ni ont a l y M e m b e r ( Depart m e nt ol C o m p u t e r S c i e n c e )

Dr. P. j j ^ - d u i b a i y . Out. ut l t’ M e m b e r ( D e p a r t m e n t o f El e c t ri c a l a n d C o m p m e r E n g i ne e r i n g ) Pr o f e s s o r J . T . . Ho we he n k o , E x t e r n a l E x a m i n e r ( D e p a r t m e n t o f El e c t ri c a l E n g i n e e r i n g , U n i v e r s i t y o f A l b e r t a , ) 0 1 ’olor A n t l i o n y W a l s h , 1992 I ' ni ve r s i t y of \ i f tori a A l l r i y n t s r e s t r e t d . T h i s d i m H u I i o n n i t t y n o t be r e p r o d u c e d i n i r h o l t o r i n p u r l , b y n i i i n e o y r n p h o r o t h e r m e a n s, t r i l l i o n ! I ht p e r m i s s i o n o f I k e a u t h o r .

(2)

S u p e r v i s o i : I

Jr .

I ) . M . M i l l e r

A b s t r a c t

S u n n!nl( il .1 m u a 11 in/ is a g e ne ra l p u r p o s e r o m b i u a t orial opt i m i z a l ion t e chi i i ( | ti e w hi c h

has heen aj;plie(| |,o m a n y ; ' 1 mis in V LS I d e s i g n . In e s s e n c e , s i m u l a t e d a n n e a l i n g is M o n t e Carl o i t e r a t i v e i m p r o v e m e n t w i t h t h e a b i l i t y t o c o n d i t i o n a l l y accept, u p ­ hill m o v e s . 'I lie n o t i o n o f a c o o l i n g s c h e d u l e is c o m m o n t o all s i m u l a t e d a n n e a l i n g i m p l e m e n t at ions. A c o o l i n g s c h e d u l e c an b e t h o u g h t o f as s i m u l a t e d a n n e a l i n g ' s coiil rol m e c h a n i s m . E x p e r i e n t i al w o rk has b e en d o n e on e s t i m a t i n g t lie c os t o f an o p t i m a l s o l u t i o n to s o m e c o m b i n a t o r i a l o p t i m i z a t i o n p r o b l e m i n s t a n c e s . S u c h an e s t i m a t e can h e u s e d to -I • l e r mi n e t e r m i n a t i o n criteria for g e n e r a l p u r p o s e o p t i m i z a t i o n t e c h n i q u e s s u c h as i t e r a t i v e i m p r o v e m e n t or s i m u l a t e d a n n e a l i n g . W e h a v e e x t e n d e d t h i s i dea a n d d e s i g n e d a c o m p l e t e s i m u l a t e d a n n e a l i n g g e n e r a l c o o l i n g s c h e d u l e b a s e d on t h e c os t o f an o p t i m a l s o l u t i o n to a p r o b l e m i n s t a n c e . W e call t h e r e s u l t a n t s c h e d u l e an t ,rh i k!< il i /nnl-din c/< il g e n e r a l c o o l i n g s c h e d u l e .

() i h • of t he m a j o r p r o b l e m s wit h s i m u l a t e d a n n e a l i n g is its l o n g c o m pu tat ion t i m e s . I’liis p r o b l e m can b e a d d r e s s e d by first u s i n g a last lieu astir t o find a g o o d initial con- ligur.it ion and t h e n a p p l y i n g s i m u l a t e d a n n e a l i n g . T h i s a p p r o a c h is c a l l e d S i m u l a t e d

S m t i r nri . I’o e x p i o l l t lie p o t e n t i a l o f s i m u l a t e d s i n t e r i n g o n e n e e d s an a p p r o p r i a t e g e n e r a l c o o l i n g s c h e d u l e . T h e e x t e n d e d g o a l - d i r e c t e d c o o l i n g s c h e d u l e is e q u a l l y a p p l i c a b l e to s i m u l a t e d a n n e a l i n g a n d s i m u l a t e d s i n t e r i n g .

ii

70

(3)

T o d a t e , no o n e c o o l i n g sc life I u U • Inis p r o v e n s u i t a b l e for all o p t i m i z a t i o n , : 1 1 -m i n s t a n c e s . In our v i e w , n o s u c h c o o l i n g s c h e d u l e e x i s t s . C o n s e q u e n t l y , w e 11 a\ e a t t e m p t e d to ' ' ' ' ' r. t h e t y p e ol p r o h l e m hest s u i t e d to o p t i m i z a t i o n by s i m u l a t e d

a n n e a l i n g a n d s i m u l a t e d s i n t e r i n g u s i n g t he e x t e n d e d goal d i re c t ed s c h e d u l e .

W e h a v e a p p l i e d t h e e x t e n d e d g o a l - d i r e c t e d s c h e d u l e to s t a n d a r d cell pla< "incut a n d f l o or pb u s n i u g p r o b l e m s u s in g b o t h s i m u l a t e d a n n e a l i n g a m i s i m u l a t e d s i n t e r i n g . W i t h i n t h i s c o n t e x t , w e h a v e c o m p a r e d t he p e r f o r m a n c e o f I he e x t e n d e d g o a l d i r e c t e d s c h e d u l e t o o t h e r p u b l i s h e d s c h e d u l e s . Our r e s ul t s i n d i c a t e that in t e r m s o f layout q u a l i t y , t h e e x t e n d e d g o a l - d i r e c t e d s c h e d u l e perl'oi m s as well or b e l t e r t h a n I lie o t h e r s c h e d u l e s . In t h i s d i s s e r t a t i o n , w e h a v e d e v e l o p e d a n e w g e n e r a l m o l i n g s c h e d u l e . O ur evul n a t i o n o f t h e e x t e n d e d g o a l - d i r e c t e c l i e d u l e s u g g e s t s that it is a i i sel ul r e s c an h c o n t r i b u t i o n in t h e a r e a o f s i m u l a t e d a n n e a l i n g algorit hms.

10

A3A

(4)

E x a m i n e r s :

Dr. D . M. Miller, S u p e r v i s o r ( D e p a r t m e n t - o f C o m p u t e r S c i e n c e )

Dr. .J.A. I ' i l li . vJ / t l J f wi t ^ ’ntal M e m b e r ( D e p a r t m e n t o f C o m p u t e r S c i e n c e )

Dr. M. S e r r a, D e p a r t m e n t a l Mei iber ( D e p a r t m e n t o f C o m p u t e r S c i e n c e )

Dr. I*'. El^yluibaly, O u t s i d e M e m b e r ( D e p a r t m e n t o f El ect ri cal a n d Oo. mput er E n g i n e e r i n g )

P r o f e s s o r J / 1 . 7 M o v v c h e n k o , E x t e r n a l E x a m i n e r ( D e p a r t m e n t o f E l e c t r i c al E n g i n e e r i n g , U n i v e r s i t y o f A l b e r t a . )

(5)

C o n te n ts

A b s t r a c t ii C o n t e n t s v L i s t o f F i g u r e s x L i s t o f T a b l e s x i i i N o t a t i o n x v A c k n o w l e d g m e n t s x v i i D e d i c a t i o n x v i i i 1 I n t r o d u c t i o n I 1.1 M o t i v a t i o n ... 1 1.1.1 S i m u l a t e d A n n e a l i n g ... I 1 . 1. 2 T h e D i g i t a l S y s t e m D es i g n I b o c e . s s ... 1 l . i . l j G o a l O r i e n t e d O p t i m i z a t i o n in V L S ' ... 4

(6)

1.2 D i s s er t a t i o n O r g a n i z a t i o n ... 5

O p t i m i z a t i o n 7 2.1 C o m b i n a t o r i a l O p t i m i z a t i o n ... $ 2 . 2 Definit ions a n d N o t a t i o n ... 9

2.9 I leu l ist ic O p l . i m i x i i l . i o i > ... 1!

2.1 ( l e n e r a l P u r p o s e Opt imi zat i o n ... 19

2. 5 M o n t i ’ Carlo O o m b i n a t o r b d O u t imi/.a,l i o n ... I I 2,ii l l er al ive I mp r o v e me nt ... I I 2. 7 A n Kmpirical E v a l u a t i o n ... IB 2.7.1 S A P N e i g h b o u r S e l e c t i o n ... 19 2 . 5 I t e r at iv e I m p r o v e m e n t K x l . e n s i o u s ... 22 2 .9 S u m m a r y ... 2-1 S i m u l a t e d A n n e a l i n g 2 5 9.1 M a c k g r o u n d ... 25 9. 2 S i m u l a t e d A n n e a l i n g M o d e l ... . . . 26 9 9 S i m u l a t e d A n n e a l i n g P r o c e d u r e ... 29 .9.1 ( l eneral ( 'ool i ng S c h e d u l e s ... 30 9.1.1 T h e C la s si c al ( l e n e r a l ( l o o t i n g S c h e d u l e ... 32 3 . 1 . 2 A S t a t i s t i c a l ( l e n e r a l C o o l i n g S c h e d u l e ... 41 3 . 5 S u m m a i y ... 46

vi

(7)

G o a I - D i r e c ( e t i A n a e a ii n g 4 8 ■1.1 A ( lon.1-Directed A p p r o a c h ... Id •1 1.1 H e r a t i v e I m p r o v e m e n t w i t h i e S i m u l a t e ! 1 A n n e a l i n g . . . 53 ■1.1.2 A n E m p i r i c a l E v a l u a t i o n ... 51 4 . 2 T h e E x t e n d e d G o a l - D i r e c t e d A p p r o a c h . . . (>l •1.2.1 I t e r a t i v e I m p r o v e m e n t w i t hi n S i m u l a t e ! 1 A n n e a l i n g . . . dd ■1.2.2 A n E m p i r i c a l E v a l u a t i o n ... 71 •1.3 S t a t i s t i c a l E v a l u a t i o n ... 76 4.'1 S u m m a r y ... S i m u l a t e d S i n t e r i n g 7C 5.1 S i m u l a t e 1 S i n t e r i n g P r o c e d u r e ... 7d 5. 2 G e n e r a ' C o o l i n g S c h e d u l e ... S I) 5. 3 A n E m p i r i c a l E v a l u a t i o n ... 81 5 . 4 S u m m a r y ... 83 I n t e g r a t e d C i r c u i t P h v s i c a l D e s i g n 8 8 6.1 T h e B i g P i c t u r e ... 88 0 .2 S t a n d a r d Cell P h y s i c a l D e s i g n ... Sd 6. 2 . 1 S t a n d a r d - C e l l P l a c e m e n t M od e l d3 6 . 2 . 2 O p t i m i z a t i o n ... d7 C.2.3 E x p e r i m e n t a l R e s u l t s ... . . . dd

(8)

fi.:j M a c r o / C u s t o m P h y s i c a l D e s i g n ... I (Jo 0.0.1 S l i c i n g F l o o r p l a n s and S l i c i n g F r e e s ... 108 ti.0.2 S . i c i n g T re e p v a l u a t i o n ... 110 0 . 0 . 3 F l oo r pl a n Mo d e l ... 11!) 0,0. 4 O p t i m i z a t i o n ... ... 1 ‘23 0 . 3 . 5 F x p e r i m o i i l a l R e s u l t s ... 120 0.4 S u m m a r y ... 102 7 C m n e l u d i n g R e m a r k s 1 3 3 7.1 ( Mo l in g S c h e d u l e S e l e c t i o n ... 100 7.2 D i s s e r t a t i o n S u m m a r y ... 107 7.0 F u t u r e W o r k ... 10!)

7.0.1 O n| , i m a l - ( ' onl i gur ut i on ( l ost E s t i m a t i o n ... 109

7 . 0 . 2 Initial T e m p e r a t u r e V a l u e lor S i m u l a t e d S i n t e r i n g ... 141 7..0.0 T i m h c r W o l f ... 141 1.0.4 M ul t i ( l o a l C o s t F u n d , i o n s for S i m u l a t e d A n n e a l i n g ... 141 7. 0 . 0 S i m u l a t e d I n v o l u t i o n ... 142 7..0.0 F i e l d ' P r o g r a m m a b l e ( l a t e A r r a y s ... 143 7.1 P p i l o g u e ... 144 B i b l i o g r a p h y 1 4 5 A Cl o n l f o r S A 1 5 0

v i i i

(9)

B E - G o a l f o r S A lfi.'S

C E - G o a l f o r S S 1 5 7

(10)

List o f F ig u res

1.1 P h a s e s o f D i g i t a l S y s t e m D e s i g n . ...

'2.1 C o s t as. ( ' o m i g m n l ion: Hills a n d V a l l e y s ...

2 . 2 ( l e n e r a l P u r p o s e ( ' o m b i n a t o r i a l O p t i m i z a t i o n ... 2 . 2 I t e r a t i v e I m p r o v e m e n t ... . . . . 2.1 h • .) S A P : O p t i m a l C o n f i g u r a t i o n . t S i a r r y ’s n e t - l i s t ) ... 2 . 5 5 x 5 S A P : N o n O p t i m a l C o n f i g u r a t i o n ( S i a t r y ’s n e t - l i s t ) . . . 2.(i S A P : O p t i m a l C o n l i g u ra t :on ( S e c h e n ’s n e t - l i s t ) ... 2.1 Cost. rs. C o n f i g u r a t i o n C u r v e ... 2. 2 S i m u l a t i n g A n n e a l i n g P s e u d o c o d e ... 2 . 2 R e l a t i o n s h i p b e t u a v n / ; a n d A t (0 < P < 0 . 9 9 ) ... 2. I P e r f o r m a n c e C r a p h ( 1 ) ... 2. 5 P e r f o r m a n c e C r a p h ( 2 ) ... 2.(i P e r f o r m a n c e C r a p h ( 2 ) ... 2 . 7 f i x 5 S A P : S i a r r y' s Net-l i st H u a n g ' s S c h e d u l e ...

x

2

12 12 ir, us 19

20

27

21

24 2 7 28 29 15

(11)

■1.1 Iterai ive I m p r o v e m e u , wi t hi n S A ... a 3

1.2 f) v 5 S A P : Siarrv'.-* Not list ( l oal S < Im■ < 1 u 1 <■... 76

l.:i Lop of T e m p o r a l > i n r . l e m p e r a ' a r e I n d e x ... AS •l.l ( ' o n l i g u r a l i o n Cost rs. T e m p o . ‘ • I n d e x ... (tl) ■1.6 S A P : Su h o p t i m a l ( ' o nl i g ur at i o n S i a r r y ’s V 1! lisl v t Act ’.*(»()) I>1 •1.6 S A P : O p t i m a l ( o u l i g n . ation S i a r r y ' s Net list ( (' os l 2 9 9 1 ... (.2

1.7 S A I L Near O p t i m a ! ( 'onl i gural ;on S e ( lien's Ne' lisl ( ( \ > \ l .’110). t»;i l.S S A P : O p t i m a l ( ' o n l i g u r a l i o n Sncl i en' s Nel lisl ( ( 'o s l 191); . . . i> I •1.9 S A P : S c l i o n ' s Net list Ooal Sel n m i l e ... ... (Pi 1.19 An P x a m p l e e - g o a l T e m p e r a t u r e ( I r a p l i ... 1)7 ■1.1 1 (o n v e r g e n c e St ra t e gy H u a n g ’s h e l i e d u l e ... 79 1.12 H o r t o n ' s ( i r a p l i ... 71 6.1 P h a s e s o f S t a n d a r d ( 'ell P h y s i c a l D e s i g n ... 99 6. 2 S k e l e t o n S t a n d a r d ( 'ell Layout . ... 91 6. 3 W i r e H o u n d i n g B o x ... 91 6.1 P l a c e m e n t A r r a y ... 9 1 6..") P l a c e m e n t A r r a y ... 9", 6 . 6 O p t i m a l Net. I / ' i igt h 1 •> s t imal i o n ... 9X 6 . 7 Q u a d r i s e c t i o n ( S t e p I,1 ... *)*} (i.cS Q u a d r i s e c t i o n ( S t e p ..2 ) ... HO 6 . 9 Q u a d r i s e c t i o n ( S t e p 3 ) ... 199

xi

(12)

(i.l() C i r c u i t .'5 P h y s i c a l R e p r e s e n t a t i o n ... 101

( i . i l Pl ac ement . Area .’<t s i i s Pl ace me nt . C lost,... 106

6 . 1 2 P h a s e s o f M a r r o / C u s t o 1’' ’’ h y s i c a l D e s i g n ... 107 6 . 1 6 S k e l e t o n M a c r o / C u s t o m La yout ... 109 0 . M Ll oor pl an a n d S l i c i n g Tr ee R e p r e s e n t a t i o n ... 110 6 . 1 6 Kloorpl an a n d S l i c i n g Tr e e R e p r e s e n t a t i o n ... I l l 6 . 1 6 A l t e r n a t i v e S l i c i n g F l o o r p l a n ... 112 6 . 1 7 A l t e r n a t i v e ' S l i c i n g T r e e R e p r e s e n t a t i o n ... 116 6. I S P lo o r pl a n w i t h Ce l l s F m b e d d e d ... 114 6 . 1 9 S l i c i n g T r e e K v a l u a t i o n ... 117 6 . 2 0 Wi r e L e n g t h E s t i m a t e ... 119 6.21 Wi r e L e n g t h E s t i m a t e ... 122 6 . 2 2 F x a m p l e F l o o r p l a n : ( ’o s t , = 9 8 7 . 5 A r e a = 8 1 0 W i r e l e n . = 9 8 7 . 6 ... 127 6 . 2 6 P r e m a t u t e C o n v e r g e n c e : C o s t = 2 0 5 7 . 5 A r e a = 1 0 6 4 W i r e l e n. = 1 8 0 4 . 5 . 128 6. 21 S A IPs Rest: C o s t = 9 6 4 A r e a = 8 7 4 W i r e l e n . = 8 9 0 ... 129 6 . 2 6 S A - l C s Rest: ( ’o s t = 9 8 9 . 6 A r e a . = 8 7 4 W i r e l e n . = 8 6 6 . 6 ... 160 6 . 2 6 S S - I I ’s Rest: C o s t = b : 2 A r e a = 8 7 4 W i r e len. 8 4 6 ... 161

x i i

(13)

L ist o f T a b les

2.1 N e i g h b o u r L a b e l i n g ... 1() 2. 2 5 x 5 S A P : I t e r at i ve I m p r o v e m e n t . ( 1 0 0 ' T r a i l s . ) ... 21 2 . 3 5 x 5 S A P : I t e r a t i v e i m p r o v e m e n t ( K n l a r g e d N e i g h b o u r h o o d ) . ( K ) O T r i a l s . ) ... 23 3.1 5 x 5 S A P : S i a r r y ’s N e t - l i s t - ( llassical S c h e d u l e ... 35 3 . 2 5 x 5 S A P : S i a r r y ’s N e t - l i s t - C l a s s ic al S c h e d u l e ... 10 3. 3 5 x 5 S A P : S i a r r y ’s Net-list, - H u a n g ’s S c h e d u l e . ( 1 0 0 T r i a l s ) ... I I 3.4 5 x 5 S A P : S o c h e n ’s N e t - l i s t - H u a n g ’s S c h e d u l e . ( 1 0 0 ' T r i a l s . ) ... Hi

4.1 5 x 5 S A P P e r f o r m a n c e S u m m a r y : S i a r r y ’s net, list. ( 1 0 0 'Trials.) . . . 55

4 .2 S A P P e r f o r m a n c e S', m m a r y : S e c h e n ’s net l i s t . ( I "0 ' T r i a l s . ) ... 55

4 .3 5 x 5 S A P P e r f o r m a n c e S u m m a r y : S i a r r y ’s net, list... 72

4.4 S A P P e r f o r m a n c e S u m m a r y : S e c l i e n ’s n e t list... 72

4 .5 1 0 x 1 0 S A P P e r f o r m a n c e S u m m a r y : S i a r r y ’s net list. (3(J 'Trials.) . . . 73

4 . G l O x 10 S A P P e r f o r m a n c e S u m m a r y : H a n d o m Wei ght net list. (3(J'I rials.) 75

(14)

4 . 8 ScIk d u l e Si r ' 1 P< i c s ... 78 5.1 . 7 x 5 S A P P e r f o r m a n c e S u m m a r y : Siarry's n e t - l i s t . ( 1 0 0 Tr i al s . ) . . 84 5 . 2 S A P P e r f o r m a n c e S u m m a r y : S e c l i e n ’s nel , - l is l . ( 1 00 T r i a l s . ) ... 84 5 . 0 1 0 x 1 0 S A P P e r f o r m a n c e S u m m a r y : S i a n y ' s n e l - l i s t . (.'i0 'Trials.) . . 85 5.4 I 0 x 10 S A P P e r f o r m a n c e S u m m a r y : R a n d o m We i ght n e t - l i st . ( 0 0 T r i al s . ) 85 5 . 5 10 - 10 S A P P e r f o r m a n c e S u m m a r y : H o r t o n ’s n e t - l i s t . ( 30 Tr i a l s . ) . , 85 5 . 0 M e a n N u m b e r o f ( ' on f ig u r a t io n s P x a m i t i e d ... 86 5 . 7 M e a n N u m b e r o f ( ' o n f ig ur at io ns P x a m i n e d ... 86 5 . 8 S c h e d u l e S t a t i s t i c s : SS M a n d S A - P ... 87 5.!) S c h e d u l e S t at is t ics: SS-I1 a n d S A - P ... 87 (i.l S t a n d a r d ( 'ell R e s u l t s ( P l a c e m e n t C o s t ) ... 102 6 . 2 S t a n d a r d - C e l l R e s u l t s ( A r e a ) ... 103 0 . 3 F l o o r p l a u n i n g P r o b l e m . ( 30 Tri al s. ) ... 125 (i. l S c h e d u l e S t a t i s t i c s : SA-11 a n d S A - P ... 126 0 .5 S c h e d u l e S t a t i s t i c s : S A - P a n d S S - l l ... 126 7.1 P x a m p l e P r o b l e m s a n d t hei r v a l u e s ... 134

x i v

(15)

N o t a t io n

C

c

a c

sc

S ( u n i r K K , T

9

a

N

m e a t i

Q

C o s l i i i n c l i o n . A v c r i i ^ i ' c o s l .

( ' l i ange ii. cos',. ( ( .s,) ( , ). A v e r a g e AC

C r e a t e s t A C for a p a r t i c u l a r t e m p e r a t u r e value. Initial t a r g e t c o n f i g u r a t i o n cost..

T a r g e t c o n f i g u r a t i o n c osl for t e m p e r a t u r e i nde x j . O p t i m a l '■onfiguration c o s t . S i m u l a t e d a n n e a l i n g c o n f i g u r a t i o n a < ( e p t a i i c e functi on. T e m p e r a t u r e d e c r e m e n t f un c t i o n . N e i g h b o u r h o o d m a p p i n g . N u m b e r of c o n f i g u r a t i o n s in a n e i g h b o u r h o o d . M e a11. P r o b l e m size. C o n f i g u r a t i o n w i t h g r e a t e s t c o s l a c c e p t e d at a, p a r t i c ul a r t e m p e r a t u r e value. C o n f i g u r a t i o n w i t h s m a l l e s t cosl ac c e p t e d at a p a r t i c u l a r t e m p e r a t u r e value1. C o n f i g u r a t i o n .s o f p a r a m e t e r s .r0 , .ip , ...

xv

(16)

s

.S |), .S1 , f t ' < i < I>) a i d 7' 7i„ 7 ' | , . . , 7;, ((> :: J < <•/) T>

T,

7 ■I'M IN ■I'MAS (.'onfiguration .space. S e q u e n c e o f a c c e p t e d c o n f i g u r a t i o n s . Initial c o nf i g u r at i on . S t a n d a r d d e v i a t i o n . T e m p e r a t u r e . S e q u e n c e o f t e m p e r a t u r e v a l u e s . Initial t e m p e r a t u r e . Fi nal t e m p e r a t u r e . T e m p e r a t u r e for e x t e n d e d g o a l - d i r e c t e d s c h e d u l e . T e r m i n a l c o n f i g u r a t i o n wi t h best c o s t . T e r m i n a l c o n f i g u r a t i o n w i t h w o r s t c o s t . x v i

(17)

A c k n o w le d g m e n ts

1 wi s h t o t h a n k m y s u p e r v i s o r , Dr. D . M . Miller, lor his help a n d g u i d a n c e d u r i n g t h e c o u r s e o f t h i s w o r k .

1 w o u l d a l s o like t o t h a n k m y c o m m i t t e e m e m b e r s , Dr. J . A . Kllis, Dr. M. Se r r a, Dr. P\ F l - G u i b a l y a n d Dr. J . T . M o w c h e i i k o .

1 a c k n o w l e d g e t h e ( i nanci al s u p p o r t from a U n i v e r s i t y o f Victoria. F e l l o w s h i p a n d 1 wi sh t o t h a n k tin* D e p a r t m e n t o f G o m p u t e r S c i e n c e for t h e u s e of its facilities.

(18)

D e d ic a tio n

To m y y ur t n l s M a r i e a n d J e r r y

l o m y imfi l l a r b a i a a n d o u r s o n , M i c h a e l J o h n P a t r i c k

(19)

C h a p te r 1

I n tr o d u c tio n

1.1

M o t i v a t i o n

M a n y p r o b l e m s in e n g i n e e r i n g a n d t h e s c i e n c e s i n v o l v e c h o o s i n g a ‘ b e s t ’ s o l u t i o n f ro m a m o n g a l a r g e n u m b e r o f possible' s o l u t i o n s . T h i s t y p e o f p r o b l e m is k n o wn a s o p t i m i z a t i o n . In t h e p a st , o p t i m i z a t i o n t o c h n k | u e s h a v e b e e n d e v e l o p e d in an nil h o c f a s h i o n y i e l d i n g a m u l t i t u d e o f s i n g l e p u r p o s e p r o c e du r e s . S o m e g e n e r al p u r p o s e t e c h n i q u e s al so e x i s t . In this d i s s e r t a t i o n w e s t u d y o n e o f t h e mo re r e c e nt ge ne r al p u r p o s e o p t i m i z a t i o n t e c h n i q u e s cal l e d S i m u l a t e d A n n e a l i n g ( a n d a. d e r i v a t i v e o f sim u l a t e d a n n e a l i n g c a l l e d S i m u l a t e d S i n t e r i n g ) .

1.1.1

S i m u l a t e d A n n e a l i n g

A l l e x i s t i n g s i m u l a t e d a n n e a l i n g i m p l e m e n t a t i o n s s t r i v e to find a ‘be st . ’ s o l u t i on to a n o p t i m i z a t i o n p r o b l e m . O n e s o l u t i o n is j u d g e d b e t t e r t ha n a n o t h e r b y s o m e me tr ic k n o w n as s o l u t i o n c o s t . A b e s t s o l u t i o n is f o u n d by s e a r c h i n g a p o r t i o n o f t h e t ot al 1

(20)

C I I M ’ T U I t I. I N T R O D U C T I O N 2 imml>cr of s o l u t i o n s t o t he o p t i m i z a t i o n p r o b l e m . Ti e e x i s t i n g s i m u l a t e d a n n e a l i n g p r o c e d u r e s t e r m i n a t e o n c e c e r t ai n t e r m i n a t i o n c ri t er i a are satisfice. A s o l u t i o n o f l e a s t c o s t e n c o u n t e r e d d u r i n g t he s e a r c h is d e e m e d t h e ‘b e s t ’ s o l i ; , i o n . T y p i c al l y , t h e s e t e r m i n a t i o n criteria are t i m e r e l a t e d a n d a r e c h o s e n w i t h o u t r e g a r d t o t h e c os t o f an o p t i m a l s o l u t i o n t o t h e o p t i m i z a t i o n p r o b l e m , b ne p r i m a r y m o t i v a t i o n for t hi s d i s s e r t a t i o n is t o d e t e r m i n e h o w t he c o s t ( e x a c t , e s t i m a t e d , or l o w e r - b o u n d ) o f an o p t i m a l s o l u t i o n can b e used t o c o n t r o l s i m u l a t e d a n n e a l i n g ’s c o n f i g u r a t i o n s e ar ch. Re s e a r c h o n s i m u l a t e d a n n e a l i n g a l g o r i t h m s c a n b e r' v i de d i n t o i h r o e a r ea s n a m e l y , a p p l i c a t i o n s , acce l erat ion and f o u n d a t i o n s . T h e a r e a o f a p p l i c a t i o n s is c o n c e r n e d wi t h a p p l y i n g t he s i m u l a t e d a n n e a l i n g t e c h n i q u e t o p r o b l e m s . R e s e a r c h in a c n d e r a l i o n is c o n c e r n e d wi t h i m p l e m e n t a t i o n issues s u c h as e f f i c i ent d a t a s t r u c t u r e s a n d efficient cost f unc t i on e v a l u a t i o n . R e s e a r c h on t h e f o u n d a t i o n s of s i m u l a t e d a n n e a l i n g is c o n ­ c e r n e d wi t h d e t e r m i n i n g w h i c h probl e m. ! are a m e n a b l e t o o p t i m i z a t i o n b y s i m u l a t e d a n n e a l i n g ; t h e d e v e l o p m e n t o f n e w c o o l i n g s c h e d u l e s an d d e t e r m i n i n g w h i c h c o o l i n g s c h e d u l e b e s t s u i t e s a p a r t i c u l a r p r o b l e m i n s t a n c e . T h i s d i s s e r t a t i o n is c o n c e r n e d w i t h r e s e ar c h o n t h e f o u n d a t i o n s o f s i m u l a t e d a n ­ n e a l i n g . In p a r t i c u l a r , w e d e v e l o p a n e w c o o l i n g s c h e d u l e c al l e d t h e e x t e n d e d goal - d i r e c t e d s c h e d u l e w h i c h is b a s e d on t h e c os t ( e x n e t , e s t i m a t e d o r l o w e r - b o u n d ) o f an o p t i m a l s o l u t i o n t o a p r o b l e m i n s t a n c e . W e g i v e g u i d e l i n e s o n w h e n t o us e t he e x t e n d e d g o a l - d i r e c t e d s c h e d u l e base d o n p r o b l e m - i n s t a n c e s t a t i s t i c s . W e a p p l y s i m u l a t e d a n n e a l i n g u s i n g t h e e x t e n d e d g o a l - d i r e c t e d s c h e d u l e to five i n s t a n c e s o f an i d e al i z ed p r o b l e m from V L S I p h y s i c a l d e s i g n . T h i s is d o n e t o e v a l u a t e t h e s c h e d u l e ’s p e r f o r m a n c e . In order t o p u t t h i s wo r k in c o n t e x t , w h a t f o l l o w s is a br i e f d i s c u s s i o n o f t h e d i gi t a l s y s t e m d e s i g n p r o c e s s o f w h i c h p h y s i c a l d e s i g n is t h e last p h a s e.

(21)

C H A P T E R I. I N T l l O D U C r i O S

R e q u i r e m e n t s

sp ec ific a tio n

b e h a v io u r a l

re p r e se n ta tio n

s tr u c tu r a l

re p r e se n ta tio n

s tr u c tu r a l

r e p r e se n ta tio n

D e s i g n

S p e c i f i c a t i o n

L o g i c

D e s i g n

C i r c u i t

D e s i g n

F u n c t i o n a l

D e s i g n

P h y s i c a l

D e s i g n

p h y s i c a l

rep re se n ta tio n

F a b r i c a t i o n

F i g u r e 1.1: Phase,s o f D i g i t a l S y s t e m D e s i g n .

(22)

( ' I I A I ’ l h . l t I. I N T H O U r c r i O S 4

1 .1.2

T h e D i g i t a l S y s t e m D e s i g n P r o c e s s

The* digit al sys ' -ri> 'l esi on p r o c e s s ( an be* d i v i d e d i n t o f i ve p h a s e s ( Fite. 1.1). In p h a s e o n e , a. (Irsiyii s p e c i f i c a t i o n is p r o d u c e d based o n t h e r e q u i r e m e n t s o f t h e s y s t e m . T h e

h-sign spe< i de at io n p h a s e is nsi i al l v a t i m e - c o n s u m i n g m a n u a l s t e p .

F u n c t i o n a l design, is c o n c e r n e d w i t h s y s t e m b e h a v i o u r . For a g i v e n s y s t e m in­ p u t , a b t l i a v i o u r a l r > pre s e n i a t i o n e n a b l e s o n e t o d e t e r m i n e t h e s y s t e m ’s o u t p u t . For e x a m p l e , a t r u t h t a b l e is o n e t y p e o f b e h a v i o u r a l r e p r e s e n t a t i o n . T h e l o p i c d c si nn p h a s e is c o n c e r n e d w i t h t h e l og i c s t r u c t u r e s t h a t i m p l e m e n t t h e f u n c t i o n a l d e s i g n . T h e c i r c u i t ( U s i y n p h a s e is c o n c e r n e d w i t h t h e t h e e le c t r i c a l c h a r a c t e r i s t i c s of ba s i c circuit e l e m e n t s . B o t h t h e s e p h a s e s c o n t r i b u t e t o a s t r u c t u r a l n p n si n l a l i o n of t h e s y s t e m w h i ch d e s c r i b e s c o m p o n e n t s of t h e s y s t e m a n d iheir i n t e r a c t i o n . In t he p h y s i c a l d c s i y n p h a s e , t h e s t r u c t u r a l a n d b e h a v i o u r a l r e p r e s e n t a t i o n s from t h e p r e v i o u s p h a s e s a re t r a n s f o r m e d t o a p h y s i c a l r e p r e s e n t a t i o n . A p h y s i c a l re pr e­ s e n t a t i o n d e f i n e s g e o m e t r i c s h a p e s u s e d in t h e f a b r i c a t i o n o f t h e d i g i t a l s y s t e m . T h e s e c o n d a r y m o t i v a t i o n for t h i s d i s s e r t a t i o n is t o g a i n e x p e r i e n c e in t h e p h y s i c a l d e s i g n o f i n t e g r a t e d c i r cu i t s . M u c h w o r k has b e e n d o n e on a u t o m a t i n g t'ne p h y s i c a l d e s i g n p h a s e . T h e i m p o r t a n c e o f t h i s p r o b l e m c o n t i n u e s t o g ro w d u e t o t h e i n c r e as i n g c o m p l e x i t y o f i n t e g r a t e d c i r c u i t s . C o n s e q u e n t l y , wo r k o n p h y s i c a l d e s i g n a u t o m a t i o n is s t il l , a n d will c o n t i n u e t o be, an a c t i v e an d c h a l l e n g i n g re searc h area.

1 .1 .3

G o a l O r i e n t e d O p t i m i z a t i o n in V L S I

O p t i m i z a t i o n in V L S I c a n b e g oa l o r i e n t e d in t h e s e n s e t h a t a c i r cu i t m u s t b e o p t i ­ m i z e d to fit i n t o a c e r t a i n area, or p e r f o r m t o s o m e s p e c i f i c a t i o n . T h i s t y p e o f s c e n a r i o

(23)

C H A P T E R I. I N T R ( ) I ) U ( ' T I ( ) N

is wel l s u i t e d t o t h e n o t io n o f u s i n g I lie est i m a l e d c o s t or lower h o u n d cost o f an o p l i m a ! s o l u t i o n to c on t r o l s i m u l a t e d a n n e a l i n g ' s s e a r c h for a host s o l u t i o n . T h i s is t h e c a s e s i n c e t h e c o s t o f an a c c e p t a b l e s o l u t i on c a n h e s u b s t i t u t e d for t h e cost o f an o p t i m a l s o l u t i o n . In t hi s wa y , s i m u l a t e d a n n e a l i n g ’s s e ar c h for a best s o l u t i o n can he t a i l o r e d for a s pe c i fi c p r o b l e m instance*.

1.2

D i s s e r t a t i o n O r g a n iza tio n

In C h a p t e r 2, w e i n t r o d u c e a f r a m e w o r k for d i s c u s s i n g o p t i m i z a t i o n . This f r a m e w o r k i n c l u d e s d e f i n i t i o n s , an o p t i m i z a t i o n b e n c h m a r k a n d a d i s c u s s i o n o f o n e o f t h e simplest, o p t i m i z a t i o i t e c h n i q u e s n a m e l y , i t e r a t i v e i m p r o v e m e n t . O p t i m i z a t i o n b y s i m u l a t e d a n n e a l i n g is i n t r o d u c e d in C h a p t e r 2 a l o n g w i t h s e ve r al c o o l i n g s c h e d u l e s ( or c on t r ol m e c h a n i s m s ) . In C h a p t e r 4 w e i n t r o d u c e t he g o a l - d i r e c t e d c o o l i n g s c h e d u l e a n d t h e e x t e n d e d g o a l - d i r e c t e d s c h e d u l e . B o t h an* b a s e d on t h e c o s t ( e x a c t , e s t i m a t e d or l o w e r b o u n d ) o f a n o p t i m a l s o l u t i o n to a p r o b l e m i n s t a n c e . T h e p e r f o r m a n c e o f t h e e x t e n d e d goal d i r e c t e d s c h e d u l e is e v a l u a t e d o n six i u s ; a n c e s o f an i d e al i z e d p l a c e m e n t p r o b l e m . C h a p t e r 5 d e s c r i b e s s i m u l a t e d s i n t e r i n g w h i c h is a n o t h e r g e n e r a l p u r p o s e o p l i m i z a t i o n t e c h n i q u e s i mi l a r t o s i m u l a t e d a n n e a l i n g . W e s h o w h o w t h e e x t e n d e d g o a l d i r e c t e d s c h e d u l e c an b e u s e d as a c o o l i n g s c h e d u l e for s i m u l a t e d s i n t e r i n g . In C h a p t e r 6, w e d e s c r i b e t w o V L S I p h y s i c a l d e s i g n s t y l e s n a m e l y , s t a n d a r d cell a n d m a c r o / c u s t o m . W e r e p o r t t h e r esul t s o b t a i n e d b y a p p l y i n g s i m u l a t e d a n n e a l i n g ( a n d s i m u l a t e d s i n t e r i n g ) u s i ng t h e e x t e n d e d g o a l - d i r e c t e d to e i g h t s t a n d a r d cell p l a c e m e n t p r o b l e m s . W e a l s o d e m o n s t r a t e h o w s i m u l a t e d a n n e a l i n g ( a n d s i m u l a t e d s i n t e r i n g ) u s i n g t h e e x t e n d e d g o a l - d i r e c t e d s c h e d u l e c an b e a p p l i e d t o m a r r o / e u s l o m

(24)

c u a i jt i;i< i. i N T i t o u r c n o N (>

placemen*..

In C h a p t e r 7, w e g i v e g u i d e l i n e s on w h e n to use the* e x t e n d e d g o a l - d i r e c t e d s c h e d ­ u l e bas ed o n p r o b l e m i n s t a n c e s t a t i s t i c s . W e c l o s e C h a p t e r 7 w i t h a d i s s e r t a t i o n s u m m a r y , i n c l u d i n g research c o n t r i b u t i o n s , and s u g g e s t i o n s for p o s s i b l e f u t u r e wo r k.

(25)

C h a p te r 2

O p tim iz a tio n

Tl i o purpose' o f t h i s c h a p t e r is t o p r o v i d e b a c k g r o u n d m a t e r i a l for ; his d i s s e r t a t i o n W e be g i n b y d e f i n i n g c o m b i n a t o r i a l o p t i m i z a t i o n . A f r a m e w o r k for d i s c u s s i n g g e n e r a l p u r p o s e c o m b i n a t o r i a l o p t i m i z a t i o n is t hen i n t r o d u c e d . T h i s f r a m e w o r k c o n s i s t s of d e f i n i t i o n s a n d n o ' a t i o n , a n d an i d e a l i z e d p l a c e m e n t p r o b l e m . T h e p u r p o s e o f t he i d e al i ze d p l a c e m e n t p r o b l e m is to pr ovi de ’ a b e n c h m a r k for c o m p a r i n g I lie p e r f o r m a m e o f g e ne r a l p u r p o s e c o m b i n a t o r i a l o p t i m i z a t i o n p r o c e dur e ’s.

I t e r a t i v e i m p r o v e m e n t is o n e o f t h e s i m p l e ’st gc i mr al p m - p e w c o m b i n a t o r i a l opl.i m i z a t i o n l e c l m i ( ] u e s . T h e l a t t e r p a r t o f this c h a p t e r c o n t a i n s a. eh’se rinl ion of i t e r at iv e1

i m p r o v e m e n t a n d an e v a l u a t i o n o f its pe’r f brmane r’ on o. i r ielealizevl placewnenl prob lerri. We’ i n c l u d e t h i s n o w b e c a u s e l a t e r in tli<* eli sser t al i em, w e descri be’ a m i e v a l u a t e o t h e r m o r e e l a b o r a t e g e ne r a l p u rpej.se c e nnbi nat or i al ejpti mi zati e m lea h n h p i e s relative*

(26)

C ! I A P T E l i 2. O P T I M I Z A T I O N

8

2.1

C o m b in a to r ia l O p tim iz a tio n

A p r o b l e m i n s t a n c e ran b e d e s c r i be d by a title, a s p e c i f i c a t i o n a n d a c pi e s t io n. For e x a m p l e , w e . an d e s c r i b e a n instance; of t wo - l e v e l B o o l e a n m i n i m i z a t i o n [8] as f ol l ows : T i t l i : T w o - l e v e l B o o l e a n M i n i m i z a t i o n . S p e d J i r a t i o n : / ( « , h) — Tib+ab ( C a n o n i c a l s u m o f p r o d u c t s e x p r e s s i o n . ) Q u e s t i on : F i n d a s u m o f p r o d u c t s r e p r e s e n t a t i o n for f ( a , b ) w h e r e t h e n u m b e r o f p r o d u c t t e r m s is m i n i m i z e d .

T h e above- problemi i ns! an e has four f e a s i b l e s o l u t i o n s n a m e l y , a b + a b , Tib+b, a b + b a m i b. All f ea s i bl e s o l u t i o n s t > a p r o b l e m i n s t a n c e s o l v e t h e g i v e n p r o b l e m .

A re*al n u m b e r is a s s o c i a t e d w i t h e a c h f easi bl e s o l u t i o n . T h i s n u m b e r is t h e c o s t e>f the* f ea s i b l e s o l u t i o n w i t h r e s p e c t to s o m e o p t i m i z a t i o n cri t er i a. A f u n c t i o n w h i c h m a p s feasible- s ol i d io n s to c o s t s is ca l l ed an o b j e c t i v e f u n c t i o n . If w e d e f i n e t h e o b j e c ­ tive- functiem in o u r e-xample- p r o b l e m t o b e t h e n u m b e r o f p r o d u c t t e r m s in a s u m o f prod net s e x p r e s s i o n t hen a b + a b , Tib+b a n d a b + b e a c h h a v e a c o s t o f 2 a n d , b h a s a cos!, ol 1.

An o p t i m a l s o l a l ion t o a p r o b l e m i n s t a n c e is a f e as i b l e s o l u t i o n w i t h m i n i m u m c o s t . W i t h r e f e r en c e lei o u r e x a m p l e p r o b l e m , t h e f e as i b l e s o l u t i o n b is an o p t i m a l solutieMi since- it h a s the- l east c o s t from a m o n g t h e f our f e a s i b l e s o l u t i o n s . A n c a r -

o p h m u l s o l u t i o n t o a p r ob l e m i n s t a n c e is a feasi bl e s o l u t i o n ivith c o s t c l o s e (in s o m e

sens(-) !i t h e cost o f an o p t i m a l s o l u t i o n .

A n o p t i m i z a t i o n p r o b l e m i n s t a n c e is c o m p l e t e l y c h a r a c t e r i z e d by a s e t o f f e a s i b l e s o l u t i o n s a n d t hei r c o st s. A n o p t i m i z a t i o n p r o b l e m i n s t a n c e is s o l v e d b y s e l e c t i n g an o p t i m a l s o l u t i o n or in s o m e c a s e s , a near o p t i m a l s o l u t i o n f ro m t h e s e t o f f e a s i b l e s o ­

(27)

C H A P T E R 2. O P T I M I Z A T I O N

9

l u t i o n s . If an o b j e c t i v e f u n c t i o n is a f u n c t i o n o f d i s c r e t e v a r i abl e s t h e n , t h e a s s o c i a t ed o p t i m i z a t i o n p r o b l e m is c a l l e d c o m b i n a t o r i a l o p l i m i z n t ion ( C O ) .

2.2

D e fin itio n s a n d N o t a t io n

Let ,s b e a c o n f i g u r a t i o n o f p a r a m e t e r s w h i c h r e p r es e n t s a fe as i bl e s o l u t i o n to an o p t i m i z a t i o n p r o b l e m i n s t a n c e. 1 Let t h e set. o f all f e as i bl e s o l u t i o n s t o a p r o b l em i n s t a n c e b e r e p r e s e n t e d by t h e s e t o f co nf ig ur a t ions .S’ (,s c .S’). T h e set ,s’ is cal l ed a

c o n f i g u r a t i o n s p a c e . Lot ( ' b e an o b j e c t i v e c o s t f u n e ' . o n . ( N o t e , deliuit ions I t hr ough

6 c o m e f rom P a p a d i m i t r i o u a n d St.eiglitz [22]. D e l i n i l i o n s 7, S a n d !) a r e pr o v i d e d by t h e a u t h o r . ) D e f i n i t i o n 1 A n i n s t a n c e o f a C O p r o b l e m is a p a i r { S , ( ' ) w h e n S is a c o n f i g u r a ­ t i o n s p a c e a n d C i s a m a p p i n g C : S -> T D e f i n i t i o n 2 A C O proble m is a s e t o f C O p r o b l e m i n s t a n c e s . D e f i n i t i o n 3 T h e a i m o f o p t i m i z a t i o n is t o Jind an o p t i m a l c o nj ig i i rn l i on »r ( S w h e r e C ( . s , ) < C ( s y ) V.s„<E,S’ T h e t e r m g l o b a l m i n i m u m c o n f i g u r a t i o n is a s y n o n y m f o r o p t i m a l c o nf i g u r a t i o n . D e f i n i t i o n 4 A n e i g h b o u r h o o d is a m a p p i n g G : S —> 2 s ( t he p o w e r s e t o f S ) d e f i n e d f o r e a c h C O p r o b l e m i n s t a n c e .

(28)

C H A T T E R 2. O P T I M I Z A T I O N

JO

D e f i n i t i o n 5 The c o n f i g u r a t i o n s y i s a n e i g h b o u r o f t he c o n f i g u r a t i o n s x ( s x , s y e s )

if

S y G ( i ( .s3.) D e f i n i t i o n 6 The c o n f i g u r a t i o n s r G S is a local m i n i m u m i f C ( s x ) < ( ' ( s y ) V s y € G ( s r ) D e f i n i t i o n 7 A d o w n - h i l l p a t h o f l e n g t h n is a s e q u e n c e ( s i , s 2, s „ J w h e r e ( a ) s r a S (I 1 ./• < n) (l>) C ( s tl) < ( '(.s / i- | ), f '(.sn- 1 ) < C'(‘s2) < (■1( ‘s l) ( c ) ‘.n G ( '"(.S,| „ I ) < •s u_ I f / ( . S n - ' i ) , 's 2 s ^ ' ( tSl ) D e f i n i t i o n 8 /I r e gi o n is a p a i r ( s x , R ) w h e r e s x G S , R C S a n d V s u R , hath s x a n d Sy exi st on the s a m e d o u m - hi l l p a t h a n d C ( s x ) < C ( s v ). D e f i n i t i o n 9 T h e c o n f i g u r a t i o n s x S is a r e g i o n a l l o c a l m i n i m u m in t h e regi on ( s x , R ) . T h r o u g h o u t , this d i s s e r t a t i o n , t h e n o t i o n s o f hi l l s a n d v a l l e y s a r e u s e d t o d e ­ s c r i b e C O p r o b l e m s . Fi g ur e 2.1 s h o w s a m o d e l c o s t ve rs us c o n f i g u r a t i o n c u r v e . In real i ty, a plot o f c o s t v e r s us c o n f i g u r a t i o n for a C O p r o b l e m i n s t a n c e w o u l d b e a m u l t i - d i n i e u s i o n a ! c o n t o u r , f o r d i s c u s s i o n p u r p o s e s , w e m o d e l t h i s c o n t o u r as a two- d i m e n s i o n a l c u r v e . In k e e p i n g wi th t hi s m o d e l , w e u s e d e s c r i p t i v e l a n g u a g e w h i c h is c o n t i n u o u s in n a t u r e . T h i s is d o n e w i t h full k n o w l e d g e t h a t c o s t v e r s u s c o n f i g u r a t i o n c o n t o u r s are d i s n e t . e .

(29)

C H A P T E R 2. O P T I M I Z A T I O N

W i t h regard t o F i g u r e 2.1, ad j ac e n t c o n f i g u r a t i o n s on t h e h o r i z o n t a l axis are n e i g h b o u r s a n d t h e c u r v e i t s e l f is v i e w e d as a terrain o f hills a n d val l eys . Kacli val l ey is c o n s i d e r e d a r eg i o n . A regi onal local m i n i m u m is l oc at e d at t h e b o t t o m of e a c h v al l e y.

A p r o c e d u r e t h a t s o l v e s a C O p r o b l e m is ca l l ed a CO a h j o n l h m . W i t h i n the a n a l o g y o f hills a n d v a l l e y s , an o p t i m i z a t i o n a l g o r i t h m c o r r e s p o n d s t o p l a c i n g a ball s o m e w h e r e on a c o s t v e r s u s c o n f i g u r a t i o n c u r v e a n d g i v i n g it t h e a b i l i t y to roll. A l g o r i t h m t e r m i n a t i o n c o r r e s p o n d s to s t o p p i n g t h e ball. T h e c o n l i g u r a l ion at whi c h t h e ball s t a r t s t o roll is c a l l e d t h e i n i t i a l c o n j u j u r a t i o n . T h e c o n l i g u r al ion with least, c o s t , e n c o u n t e r e d w h i l e t h e ball is rol l i ng, is c al l ed t he t e r m i n a l conjii/u r a t i o n .

2 .3

H e u r is tic O p tim iz a tio n

A n o p t i m i z a t i o n a l g o r i t h m is e x a c t if it, y i e l d s an o p t i m a l con l i g u r a l , i o n to e v e r y i n s t a n c e o f an o p t i m i z a t i o n p r o b l e m . O n e o b v i o u s e x a c t a l g o r i t h m for C O is to e x a m i n e e a c h c o n f i g u r a t i o n in t h e c o n f i g u r a t i o n s p a c e o f t he p r o b l e m i n s t a n c e and s e l e c t a c o n f i g u r a t i o n w h i c h m i n i m i z e s t h e o b j e c t i v e func t i on.

M a n y C O p r o b l e m s a r e N P - h a r d [33], All k n o w n exact, a l g o r i t h m s for N P hard o p t i m i z a t i o n p r o b l e m s r e qu i r e a c o m p u t a t i o n a l ef f ort w h o s e w o r s t c a s e t i m e c o m p l e x i t y is a t l e a s t e x p o n e n t i a l in t h e s i z e of t h e p r o b l e m i n s t a n c e . For l a r g e i n s t a n c e s of N P - h a r d o p t i m i z a t i o n p r o b l e m s , t h e use o f a n y k n o w n exact, o p t i m i z a t i o n a l g o r i t h m w o u l d b e i m p r a c t i c a l s i n c e it w o u l d requi re p r o h i b i t i v e l y l arge e x e c u t i o n t imes.

A s an a l t e r n a t i v e t o e x a c t a l g o r i t h m s , h e u r i s t i c or a p p r o x i m a t e a l g o r i t h m s ma y b e used t o s o l v e C O p r o b l e m s . H e u r i s t i c a l g o r i t h m s yi el d o p t i m a l or ne ar o p t i m a l c o n f i g u r a t i o n s t o N P - h a r d o p t i m i z a t i o n p r o b l e m s a n d t y p i c a l l y r e q u i r e a r o mput a.

(30)

C I I A I ’ T I M 2. O P T I M I Z A T I O N

Cost

Configurations

(31)

C H A P T E R 2. O P T I M I Z A T I O N !:t

p r o c e d u r e g e n e r a l p u r p o s e c o m b i n a t o r i a l o p t i m i z a l ion(i n .s,,; out .s(,); var .s0,.s;), . s, . Y:eonl i gurati on;

(* «s0 = init ial c o n f i g u r a t i o n *) (* tip — t e r mi n a l c o n f i g u r a t i o n * ) (* X --= current c o nf i g u r at i on * ) (* ,s = s e l e c t e d n e i g h b o u r e o n l i g u r a l i o n *) b e g i n •V : = .s0 ; r e pe at select, .s; ('*’ ,s t <7( A ) *) if a e e e p t ( d ' ( . s ) ) t hen A ' : = el s e N O P u n t i l t e r m i n a t i o n V = A ' e n d ; F i g u r e 2.2: G e n e r a l P u r p o s e C o m b i n a t o r i a l O p t i m i z a t i o n .

t i o n a ! effort w h o s e wors t - c a. se t i m e c o m p l e x i t y is b o u n d e d p o l y n o m i a l y in I lie s i ze of t h e p r o b l e m i n s t a n c e , b o w o r de r p o l y n o m i a l t i m e a l g o r i t h m s c a n be pra< tiea.1.

2 .4

G en era l P u r p o s e O p tim iz a tio n

O p t i m i z a t i o n a l g o r i t h m s w h i c h are p r o b l e m i n d e p e n d e n t ar e cal l ed t / n i c r a l p u i y w s t . G e n e r a l p u r p o s e o p t i m i z a t i o n algorit h m s p r e s u p p o s e t h e e x i s t e n c e o f a, c u r r e n t con f i g u r a t i o n , a n e i g h b o u r s e l e c t i o n m e c h a n i s m a n d an o b j e c t i v e f i i m t i o n . W e l i mi t o u r a t t e n t i o n to a l g o r i t h m s w h i c h p r o c e e d by s e l e c t i n g a n e i g h b o u r o f t h e c u r r e n t c on f ig

(32)

( H I A T I E H 2. O P T I M I Z A T I O N

14

u r a t i o n a n d a c c i p l i n g or r e j e c t i n g it ba s e d on its c o s t , i n the* e v e n t t h a t t h e s e l e c t e d n e i g h b o u r c o n f i g u r a t i o n is a c c e p t e d , t h e n it is m a d e t h e c ur r e n t c o n f i g u r a t i o n . If t h e s e l e c t e d n e i g h b o u r c o n f i g u r a t i o n is r ej ecte d t h e n it is d i s c a r d e d . T h e a b o v e p r o c e d u r e is r e p e a t e d unt i l s o m e t e r m i n a t i o n c ri t e r i a are s at is f ie d. A p s e u d o c o d e d e f i n i t i o n of g e n e r a l p u r p o s e c o m b i n a t o r i a l o p t i m i z a t i o n is s h o w n in F i g u r e 2. 2.

2.5

M o n t e C arlo C o m b in a to r ia l O p tim iz a tio n

T h e term M o n t e ''at'lo Method, is u s e d t o d e s c r i b e a n y a l g o r i t h m t h a t e m p l o y s r a n d o m n u m b e r s [|.r)j. C o m b i n a t o r i a l o p t i m i z a t i o n u s i n g a M o n t e C a r l o m e t h o d is cal l ed

M o n t i C u r i o c o m b i n a t o r i a l o p t i m i z a t i o n . M o n t e C a r l o c o m b i n a t o r i a l o p t i m i z a t i o n t e c h n i q u e s c o m e in m a n y d i ff e r e n t f la v o u r s [20]. In t h i s d i s s e r t a t i o n , w e e x a m i n e four M o n t e C a r l o c o m b i n a t o r i a l o p t i m i z a t i o n t e c h n i q u e s n a m e l y , M o n t e C a r l o I t e r a t i v e

I m p r o i u m e n ! ( C h a p t e r 2) , S i m u l a t e d A n n e a l i n g , S i m u l a t e d S i n t e r i n g ( C h a p t e r s 3,4

a n d fi) and S i m u l a t e d Invol ut i on ( C h a p t e r 7).

2 .6

I te r a t iv e I m p r o v e m e n t

I t e r a t i v e i mp r ov e me nt , is a g e n e r a l p u r p o s e o p t i m i z a t i o n t e c h n i q u e . S t a r t i n g f r om a n initial c o n f i g u r a t i o n .s0 £ S , i t e r a t i v e i m p r o v e m e n t t r a n s f o r m s s0 t o a t e r mi n a l c o n f i g u r a t i o n .s;, £ .S', d his is a c h i e v e d by c o n s t r u c t i n g a s e q u e n c e o f c o n f i g u r a ­ t i o n s ,sn , ,'<|, . . . ,s . s y ' f l i e c o n f i g u r a t i o n .s, m u s t c o m e f ro m s , _ i ’s n e i g h b o u r h o o d ( 1 b ' S !>)■ In a d d i t i o n , C (•*>’,) < Cl'(.s,_t ), ( 1 < i < p).

(33)

C H A P T E R

2.

O P T I M I Z A T I O N p r o c e du r e ' t e r a l i v e i m p r o v e m e n t fin ,s0 ; out s,,); var s0,.s,,,.s,.Yu'oidiguration; begi n A : = .so; re pe at s e l e c t .s; (* ,s r; f i ( A ’ ) *) i l ' C( . s) < ( ' ( A’ ) t h e n X : = .s unt il H A " ) < f ( . s , . ) V . s , 6 t.’(A'); .s(J: = X (’iu l ; F i gu r e 2.U: I t e r a t i v e I m p r o v e m e n t . T h e a c c e p t a n c e ol s, as t h e s u c c e s s o r t o .s,_| is s a i d t o he a d o m i - h i l l mot>t .since t h e c o s t ol’ Si m u s t h e less t h a n t he c o s t o f . s,_( (I < i < j>)T I t e r a t i v e i m p r o v e m e n t p e r m i t s o n l y cl own-hi 11 m o v e s a n d t e r m i n a t e s w h e n n o f urther d o w n hill m o v e s are p o s s i b l e , i . e . , w h e n C'(sp) < ( ' ( s / . ) V.s* P f / ( s /t). A p s e u d o c o d e d e f i n i t i o n o f i t e r a t i v e i m p r o v e m e n t is s h o w n in F i g u r e 2.11.

T o d a t e , it has not been p o s s i b l e t o g i v e a useful u p p e r h o u n d on i t e r a t i v e im p r o v e m e n t ’s w o r s t c as e t i m e c o m p l e x i f y [36]. N e v e r t h e l e s s , i t e r a t i v e i m p r o v e m e n t h a s b e e n s h o w n t o be e f f e ct i ve for s o l v i n g m a n y di f fe r e n t c o m b i n a t o r i a l o p t i m i z a t i o n p r o b l e m s [1 1, 16]. In t h e a b s e n c e o f a n y m e a n i n g f u l t h e o r e t i c a l a n a l y s i s , w e t u r n our a t t e n t i o n t o a n e m p i r i c a l s t ud y.

2 N o t e t h a t C ( s t ) m u s t Ik* s t r i c t l y less t h a n < 7 ( » , - i ) , e q u a l i t y rloes not c o n s t i t u t e a d o w n - l u l l

(34)

C U A l ' T m . 2. O P T I M I ' / A T I O N

1 6

2 .7

A n E m p irica l E v a lu a tio n

A s u i t a b l e p r o b l e m is required for e v a l u a t i n g o p t i m i z a t i o n p r o c e d u r e s . W e c h o o s e s q u a r e a r r a y p l a c e m e n t [29] as a b e n c h m a r k s i n c e it is r e p r e s e n t a t i v e o f p l a c e m e n t in VL S I p h y s i c a l d e s i gn . A s i n g l e a p p l i c a t i o n o f an o p t i m i z a t i o n p r o c e d u r e t o a p r o b l e m i n s t a n c e is c al l ed a. tri al . W e e v a l u a t e an o p t i m i z a t i o n p r o c e d u r e by p e r f o r m i n g m u l t i p l e trials o n a g i v e n p r o b l e m i n s t a n c e .

Mach trial s t a r t s wi t h a di f f e r e nt i ni t i al c o n f i g u r a t i o n . Let us s a y t h a t w e p e r f o r m

1 0 0 trials. T h i s w o u l d result in 100 p o t e n t i a l l y di f f e r e n t t e r m i n a l c o n f i g u r a t i o n s . W e record t e r m i n a l - c o n f i g u r a t i o n c o s t s t a t i s t i c s t o m e a s u r e a p r o c e d u r e ’s a b i l i t y t o p r o d u c e , o n a v e r a g e , q u a l i t y t e r m i n a l c o n f i g u r a t i o n s for a g i v e n p r o b l e m i n s t a n c e . W e al so record t h e a v e r a g e n u m b e r o f c o n f i g u r a t i o n s e x a m i n e d o v e r all t r i al s for a g i v e n p r o b l e m i ns t anc e 1. T h i s s t a t i s t i c is a m e a s u r e o f h o w l o n g a p r o c e d u r e t a k e s , o n a v e r a g e , t o l o c a t e a t e r m i n a l c o n f i g u r a t i o n . K n o w l e d g e o f t h e s e s t a t i s t i c s is i m p o r t a n t for t h e p u r p o s e o f c o m p a r i n g a n d c o n t r a s t i n g d i ff e r en t p r o c e d u r e s . For e x a m p l e , c o n s i d e r t h e s i t u a t i o n w h e r e w e c o m p a r e t w o o p t i m i z a t i o n p r o c e d u r e s a n d find that t h e a v e r a g e t e r m i n a l c o n f i g u r a t i o n cost for p r o c e d u r e 1 is less t h a t t h e ave r age1 t e r m i n a l c o n f i g u r a t i o n c o s t for p r o c e d u r e 2. W e u s e t h i s as an i n d i c a t i o n t h a t p r o c e d u r e 1 is m o r e likely t o p r o d u c e a t e r m i n a l c o n f i g u r a t i o n w i t h o v e r a l l l ea s t c o s t . T h i s is n o t t o s a y t h a t for a g i v e n n u m b e r o f t r i al s , p r o c e d u r e 1 will a l w a y s p r o d u c e a t e r m i n a l c o n f i g u r a t i o n w i t h overal l l east c o s t .

Srpiare a r r a y p l a c e m e n t ( S A P ) is an i deal i z e d f o r m o f f l o o r p l a n n i n g . S A P is a g e o m e t r i c p r o b l e m c o m . ‘m o d w i t h p l a c i n g u n i t - s q u a r e o b j e c t s in t h e p l a n e . T h e s e u n i t - s q u a r e o b j e c t s are cal l ed cells.

(35)

C H A P T E R 2. O P T I M I Z A T I O N

l o c a t i o n c a n a c c o m m o d a t e a s i n g l e coll. T l i o t o t al n u m b e r o f colls l o Ik* p l a c ed in t h e a r r a y is m 2. ( ’oils a r e l a be l e d 1 t o i n 2.

Lot l\I b o t h e s e t of cell l ab e ls for t h e S A P p r o b l e m . A / tro-t<■nninnl n v t - l m t is a s u b s e t IT o f t he ( ’a r t e s i an p r o d u c t o f M w i t h itself. An e l e m e n t of IP is c al l e d a n r / . A no* r e p r e s e n t s a c o n n e c t i o n b e t w e e n t w o cells. In t e r m s o f VI , SI p h y s i c a l d e s i g n , a net is r e a l i z e d by a wi re in a n i n t e g r a t e d ci r c ui t . A S A P p r o b l e m c o n f i g u r a t i o n is an a s s i g n m e n t of cel' . t o cel! l o c a t i o n s . T h e c o n f i g u r a t i o n is a n t i g h b o u r o f s.,. ( s , . , . s v c; .S’) if ,Sj. c a n be t r a n s f o r m e d t o s„ by s w a p p i n g t he p o s i t i o n s o f t w o c el l s in .s,.. C o n s e q u e n t l y , each c i . n l i g ur a l i o u has ( ^ = N n e i g h b o u r s a n d |A'| = ( m 2)!.

t 2 /

For a g i v e n c o n f i g u r a t i o n , e a c h cell has a l o c a t i o n w h i c h w e d e s c r i b e b y t h e a n ay c o o r d i n a t e s ( . r , y ) (1 < . r, y < i n ) . F a e h net, is g i v e n a c o s t w h i c h c o r r e s p o n d s t o t h e s u m o f t h e h o r i z o n t a l d i s t a n c e a n d ve r t i c al d i s t a n c e b e t w e e n t h e n e t ’s ( el ls . For e x a m p l e , if a net c o n n e c t s cell 1 l o c a t e d at, (3, b) t o cell 2 l o c a t e d at ( b , l ) t h e n t h e ve r t i c al d i s t a n c e b e t w e e n cells 1 a n d 2 is 3 a n d t h e h o r i z o n t a l d i s t a n c e bet we e n c el l s 1 a n d 2 is d. T h e s u m o f t h e h o r i z o n t a l d i s t a n c e a n d t h e v e r t i c a l d i s t a n c e is 7 a n d is c a l l e d t h e M a n h a t t a n d i s t a n c e . T h e M a n h a t t a n d i s t a n c e for e a c h n e t is c a l c u l a t e d a n d s u m m e d t o g i v e t h e t o t a l M a n h a t t a n d i s t a n c e . T h e t ot al c o s t o f a c o n f i g u r a t i o n is d e f i n e d as l i ve t i m e s t h e t o t a l M a n h a t t a n d i s t a n c e [29]. T h e f ac t or five is o f S i a r r y ’s i n v e n t i o n . It h a s b e e n carri ed o v e r i n t o o u r work s i m p l y b e c a u s e w e w a n t t o u s e t h e s a m e c o s t f u n c t i o n as S i a r r y u s e d in his tri al s. T h e o b j e c t i v e o f t h e HAP p r o b l e m is t o d e t e r m i n e an a s s i g n m e n t of c el l s t o cell he c a t i o n s s u c h t h a t t h e tot al M a n h a t t a n distance* b e t w e e n c o n n e c t e d cells is m i n i m i z e d .

(36)

C l i A P T K H 2. O P T I M I Z A T I O N

18

10 15 1 20 25

L

e 0 M1 If1 24 23 22 21 i

fl

in !» i 7

r q

17 1 r.

J

11 10 L -

..

Figure' 2 . <1: 5 x 5 S A P : O p t i m a l C o n f i g u r a t i o n . ( S i a r r y ’s n e t - l i s t ) . a n d is e-emstructeel s u c h t h a t in an o p t i m a l p l a c e m e n t , e a c h n e t is c o n n e c t e d t o its ne a r e s t h o r i z o n t a l a n d verti cal n e i g h b o u r s only. A n o p t i m a l s o l u t i o n h a s c o s t :

5 * 2 ( m " — rn)

F i g u r e 2 . 1 s h o w s an o p t i m a l 5 x 5 p l a c e m e n t ( n e a r e s t n e i g h b o u r s c o n n e c t e d ) . For c o m p a r i s o n purpose's, F i g u r e 2. 5 s h o w s a n o n - o p t i m a l p l a c e m e n t . E a c h c o n f i g u r a t i o n h a s 800 n e i g h b o u r s a n d |.S’| = 25! ~ 102S. T h e r e are 8 u n i q u e o p t i m a l c o n f i g u r a t i o n s . Fae'h e>ptimal c o n f i g u r a t i o n has a. c o s t o f 200.

T'he' seronel n e t - l i s t a p p l i e s o n l y to 5 x 5 S A P a n d c o m e s f r o m S e c h e n [27]. F i g u r e 2 . 0 s h o w s an o p t i m a l p l a c e m e n t . T h e r e are 16 u n i q u e o p t i m a l c o n f i g u r a t i o n s . E a c h ei pt i mal e-onl i gurati on h a s a c o s t o f 190.

(37)

C H A P T E R 2. O P T I M I Z A T I O N

I!)

F i g u r e 2.5: 5 x 5 S A P : N o n O p t i m a l ( 'onfiguratiori ( S i a r r y ’s net list.).

N e i g h b o u r L a b e l 1 2 3 V I 2 ... 1 i n 2 i n 2 -f 1 Ce l l Pai r 1 , 2 i ; i l / l 1 , n i 2 '> :> 2 / 1 ;V m 2 I r!’a l ) l e2.1: N e i g h b o u r La b e l i n g

2.7 .1

S A P N e i g h b o u r S e l e c t i o n

T h e e s s e n c e ol' a n y n e i g h b o u r select,ion m e c h a n i s m is I,he o r d e r in w h i c h n e i g h b o ur c o n f i g u r a t i o n s a ,-e e x a m i n e d . For e a s " o f r ef e r e n c e , let T a b l e 2 . 1 d e f i n e a n e i g h b o u r l a b e l i n g for t h e j n x r n S A P p r o b l e m , t . t j . , t h e c o n f i gu r a t i on f ound by s w a p p i n g t he p o s i t i o n s o f cell 1 a n d cell 2 is c al l ed n e i g h b o u r 1. N e i g h b o u r s e l e c t i o n correspond:; to s e l e c t i n g a n e i g h b o u r from t h e c ur r e n t c o n f i g u r a t i o n ' s n e i g h b o u r h o o d . M a n y different n e i g h b o u r o r d e r i n g s a r e p o s s i b l e .

(38)

c i i A i n ' m 2 . o r r i M i Z A T i O N

20

2i 20 Fi gur e 2.0: S A P : O p l i m a l C o n f i g u r a t i o n ( S e c h e n ’s net-list,). W e use t h e f o l l o w i n g p r oc e d u r e . S e l e c t i o n f rom t h e n e i g h b o u r h o o d is p e r f o r m e d in a p s e u d o r a n d o m order. A p s e u d o - r a n d o m n u m b e r in t h e r a n g e [1,A/] is g e n e r a t e d . If t h e n u m b e r is s a y 5, t h e n n e i g h b o u r 5 is s e l e c t e d . O n c e a n e i g h b o u r has b e e n s e l e c t e d , it is not c o n s i d e r e d agai n for s e l e c t i o n unt i l s u c h t i m e a s all o t h e r n e i g h b o u r s h a v e been s e l e c t e d . N e i g h b o u r s e l e c t i o n c o n t i n u e s by g e n e r a t i n g a d d i t i o n a l p s e u d o ­ r a n d o m n u m b e r s . N o t e t h a t for i t e r a t i v e i m p r o v e m e n t , t h e o r d e r in w h i c h n e i g h b o u r s are e x a m i n e d is u s u a l l y fixed. It e r at i ve i m p r o v e m e n t is s o m e t i m e s referred t o as

M o u l t C a r l o i t e r a t i v e i m p r o v e m e n t w h e n n e i g h b o u r h o o d e x a m i n a t i o n is p e r f o r m e d

in a p s e u d o r a n d o m f as hi on.

W e e v a l u a t e d i t e r a t i v e i m p r o v e m e n t o n t h e 5 x 5 S A P p r o b l e m . T h e r e s ul ts a r e s h o w n in T a b l e 2.2. For b o t h not - l i st s , 100 tri al s w e r e p e r f o r m e d . E a c h trial h a d a di f f e r e nt s t a r t i n g c o n f i g u r a t i o n w h i c h wa s c h o s e n a t r a n d o m . In t h e c o s t a n d c o n

Referenties

GERELATEERDE DOCUMENTEN

It is realistic to expect that an effective road safety policy in CEECs will result in a smaller decrease in road safety, as was the case in highly

Alhoewel het hier om een orematiegra.Wcld gaat bestond er toch een zekere verscheidenheid iÏ!n het · begrail'eni!'lr1Jtueel. 1n de andere kuilen waren de verbrande

In het Rhenodunale district (Wassenaar) vertoont suppletiezand dat over de zeereep heen gestoven is, ten opzichte van niet overstoven ‘natuurlijk’ duinzand, significant

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:.. • A submitted manuscript is

†, Include the value of the outstanding loan in terms of section 3(2) of the Estate Duty Act; ‡, Disregarding the outstanding loan as an asset in the disposer’s deceased estate

The current thyristor based electric braking circuit, used during speed control for grid synchronisation, can be used to limit the output power, however this circuit is not

Dankzij de medewerking van de Dienst Bouwprojecten en het aannemersbedrijf SDH was het voor de Dienst Stadsarcheologie van de Stad Gent mogelijk een beperkt archeologisch

daarbij aandacht voor de zelf- en samenredzaamheid van de cliënt en diens naastbetrokkenen en houdt P2-K1: Voert ondersteunende werkzaamheden uit Verantwoordelijkheid en