Generalized Runge Kutta Method

download Generalized Runge Kutta Method

of 14

Transcript of Generalized Runge Kutta Method

  • 7/25/2019 Generalized Runge Kutta Method

    1/14

    Num er. M ath. 33, 5 5- 6 8 (1979)

    Numerische

    athemaUk

    9 by Springer-Verlag 1979

    General ized Runge Kutta Methods

    of Order Four with Stepsize Control

    for Stiff Ordinary Differential Equations

    1 P e t e r K a p s , 2 P e t e r R e n t r o p

    x Institut fiir Math em atik I, Universit~it Inns bruck , Tech nikerstr. 13, A -6020 Innsb ruck , Austria

    z Institut fiir Ma them atik der Techn ischen Un iversit~t Miinchen,

    Arcisstr. 21, D-800 0 Miinchen 2, Germ any (Fed . Rep.)

    S u m m a r y . G e n e r a l i z e d A (c~ )-stab le R u n g e - K u t t a m e t h o d s o f o r d e r f o u r w i t h

    s t ep s i ze c o n t r o l a r e s t u d ie d . T h e e q u a t i o n s o f c o n d i t i o n f o r t h is c l a ss o f s e m i -

    i m p l i c i t m e t h o d s a r e s o l v e d t a k i n g t h e t r u n c a t i o n e r r o r i n t o c o n s i d e r a t i o n .

    F o r a p p l i c a t io n a n A - s t ab l e a n d a n A ( 8 9 .3 ~ m e t h o d w i t h s m a ll

    t r u n c a t i o n e r r o r a r e p r o p o s e d a n d t e s t re s u l ts f o r 25 s ti ff i n i ti a l v a l u e

    p r o b l e m s f o r d if f er e n t to l e r a n c e s a r e d i s c u ss e d .

    Subjec t c lass i f ica t ions .

    A M S ( M O S ) : 6 5 L 0 5 ; C R : 5 . 1 7 .

    1. Introduct ion

    I n i t i a l v a l u e p r o b l e m s w i t h s t r o n g l y d e c r e a s i n g a n d i n c r e a s i n g s o l u t i o n c o m p o -

    n e n t s a r e c a l l e d s ti ff p r o b l e m s . T h e y m a i n l y a p p e a r i n c h e m i c a l k i n e ti c s, e l e c tr i c

    c i r c u i t s a n d c o n t r o l t h e o r y . U s u a l i n t e g r a t i o n r o u t i n e s a s c o m p a r e d i n D i e k h o f f

    e t a l. [ 6 ] , E n r i g h t e t a l . [ 7 ] f ai l, b e c a u s e o f t h e d i f f e r e n t g r o w t h o f t h e s o l u t i o n

    c o m p o n e n t s . N e w s t ab i li ty r e q u i r e m e n t s l ik e A - s ta b i li ty h a v e b e e n i n t r o d u c e d t o

    o v e r c o m e th e s e p r o b l e m s , s ee D a h l q u i s t [-5 ], G r i g o r i e f f [ 1 0 ] .

    T h e p r e s e n t r e p o r t is c o n c e r n e d w i th g e n e r a li z e d R u n g e - K u t t a m e t h o d s o f

    o r d e r f o u r w i t h t h r e e f u n c t i o n e v a l u a t i o n s p e r s t e p . A s t e p s i z e c o n t r o l i s

    i m p l e m e n t e d b y e m b e d d i n g a t h ir d o r d e r m e t h o d . O n e e v a l u a ti o n o f t h e Ja c o b i

    m a t r ix a n d t h e s o l u t i o n o f a l i n e a r e q u a t i o n s y s t e m o f o r d e r n is n e c e s s a r y p e r

    s te p . A n A - s t a b l e a n d a n A ( 8 9 . 3 ~ a l g o r i t h m a r e t e s te d b y s o l v i n g 2 5 s ti ff

    i n i t i a l v a l u e p r o b l e m s f r o m B e d e t , E n r i g h t , H u l l [ 2 ] a n d E n r i g h t , H u l l , L i n d b e r g

    [ 8 ]

    2 . G e n e r a l i z e d R u n g e K u t t a M e t h o d s

    R O W - M e t h o d s .

    T h e a u t o n o m o u s i ni t ia l v al u e p r o b l e m :

    y ( x ) = f ( y ( x ) ) ,

    Y( Xo) Yo

    (2.1)

    0 0 2 9 - 5 9 9 X / 7 9 / 0 0 3 3 / 0 0 5 5 / $ 0 2 . 8 0

  • 7/25/2019 Generalized Runge Kutta Method

    2/14

    56 P. K aps and P. Re ntrop

    i s c o n s i d e r e d i n a n - d i m e n s i o n a l r e a l o r c o m p l e x s p a c e . A n u m e r i c a l s o l u t i o n o f

    t h e f o l l o w i n g t y p e i s s t u d i e d :

    yh(Xo + h)

    =Y 0 + ~

    ci ki

    (2.2)

    i=

    ( I - ; ~ h f ' ( Y o ) ) k i = h Y o + ~ c ~jk j + h f ' ( y o ) ~ 7 ij kj i = l , . . . , s .

    j = l j = l

    Th e co e f f i c i en t s 7 , ~ ij, ~ ij a r e r ea l n u m b e r s , h d e n o te s t h e s t ep s i ze ,

    f ' (Y o )

    t h e

    J a c o b i - , I t h e n x n i d e n t i t y m a t r i x a n d s t h e n u m b e r o f st a ge s . T h e v e c t o r s k~ (i

    = 1 , . . . , s ) a r e c o m p u t e d b y s o l v i n g a s y s te m o f l i n e a r e q u a t i o n s o f o r d e r n fo r s

    d i f f e r e n t r i g h t h a n d s i d e s .

    M e t h o d ( 2 . 2 ) i s c a l l e d R o s e n b r o c k - W a n n e r m e t h o d , s h o r t R O W - m e t h o d .

    T h e f i r s t w h o s e e m e d t o h a v e s t u d i e d s i m i l a r f o r m u l a s w a s R o s e n b r o c k [ 1 8 ] .

    W a n n e r [ 2 0 ] i n t r o d u c e d t h e c o ef fi ci en ts ~ u a n d p r o p o s e d t h e t h e o r y o f B u t c h e r

    s e ri e s [1 1 ] f o r d e r i v a t i o n o f t h e e q u a t i o n s o f c o n d i t i o n . I n [ 1 3 ] a n d W o l f b r a n d t

    [ 2 1 ] t h e s e m e t h o d s a r e c a l l e d m o d i f i e d R o s e n b r o c k m e t h o d s , i n N o r s e t t ,

    W o l f b r a n d t [ 1 7 ] R O W - m e t h o d s .

    F o r 7 = 7 u = 0 t he R O W - m e t h o d s r ed u c e t o u su al R u n g e - K u t t a m et ho d s.

    T h e r e f o r e R O W - m e t h o d s c a n b e c o n si d e re d a s g e n er a li ze d R u n g e - K u t t a m e t h-

    o d s .

    Stab i l i ty P roper ties o f RO W -M ethods . T o s t u d y t h e s ta b il it y p r o p e r t i e s o f R O W -

    m e t h o d s , t h e s c a l a r t e s t d i f f e r e n t i a l e q u a t i o n i s u s e d :

    y ' = 2 y , y ( X o )= Y o ; 2 ~ l U , y o O l 2 , y : I R - ~ C

    S i nc e f ' ( y ) = 2 , it h o ld s : k~ =R z( z ) yo , z=2h , w h e r e Ri(z ) a r e r a t i o n a l f u n c t i o n s

    w i th d e n o m i n a t o r ( 1 - 7 z ) ~ a n d d e g r e e o f n u m e r a t o r < i . T h u s t he n u m e r i ca l

    s o l u t i o n

    Yh

    is :

    Yn = R (z) Yo

    (2.3)

    ~,, - , , P (z )

    w i t h t h e s t a b i l i t y f u n c t i o n : R ( z ) = 1 +

    ~= 1 Q ~ A z J - ~ z )

    F o r a r a t i o n a l a p p r o x i -

    m a t i o n (2 .3 ) o f o r d e r p h o l d s :

    P r o p o s i t i o n ( 2 .4 ) . T h e s t a b i l i t y f u n c t i o n o f a R O W - m e t h o d w i t h o r d e r p >__s is

    g i v e n b y

    R ( z ) = (1 - y z)- - - ~ k= o

    w h e r e

    ( n + ~ ] x '

    / J ~ , ) ( x ) = i ~ 0 ( - 1 ) i \ n - i i ~ -

  • 7/25/2019 Generalized Runge Kutta Method

    3/14

    G e n e r a l i z e d R u n g e K u t t a M e t h o d s 5 7

    s t a n ds f o r t h e g e n e r a l iz e d L a g u e r r e p o l y n o m i a l s , s e e A b r a m o w i t z , S t e g u n [ 1 ] .

    R z) is a r a t i o n a l a p p r o x i m a t i o n t o e ~ o f o r d e r > s .

    Proof

    A p p l y i n g T h e o r e m 4 o r P r o p o s i t i o n 6 o f N o r s e t t , W a n n e r [ 1 6 ] o n e c a n

    s h o w t h a t P z) i s d e t e r m i n e d u n i q u e l y b y Q z ) . I n o u r c a s e t h e s y m m e t r i c

    p o l y n o m i a l s S , a r e :

    S , = ( ~ ) 7 ' . I t f o l l o w s :

    e z)= s

    k=o i=o k i

    B y m e a n s o f t h e s t a b i li t y f u n c t i o n o n e c a n c h a r a c t e r i z e s o m e s t a b i li t y p r o p e r t i e s

    v e r y c o n v e n i e n t l y .

    On e h as s t ab i l i t y a t i n f in i ty , i f f :

    l i m I R ( z ) ] = L s 1 ) < 1 , w h e r e L s : = / 3 ~ (2 .5 )

    Z ~ O 9

    \ y ]

    F or y > 0 a m e th o d (2.2 ) is A-s tab le , i f f:

    I R i y ) l< l fo r ye lR (2.6 )

    o r e q u i v a l e n t l y , i ff t h e E - p o l y n o m i a l ( s e e N o r s e t t [ 1 5 ] ) s a t is f ie s :

    E y )=lQ i y )12-1P i y ) lZ>O

    Vy e lR. ( 2 .7 )

    Embedded ROW -Methods . E r r o r e s t i m a t i o n a n d s t e p s i z e c o n t r o l i s p e r f o r m e d

    u sin g t w o e m b e d d e d m e t h o ds . A R O W - m e t h o d o f o r d e r 4 :

    Yh Xo + h) = Yo + ~ ci ki (2.8)

    i = 1

    a nd a R O W - m e t h o d o f o r d e r 3:

    Y h X o + h ) = y o + ~ ~ i k i g < s)

    i = 1

    a r e c o m b i n e d , w h e r e t h e c o e f f i c ie n t s 7, Y , j, cq~ ( i = l , . . . , s , j = l , . . . , i - 1 ) a n d

    t h e r e f o r e t h e k , a r e t h e s a m e f o r b o t h f o r m u l as . T h e r e s u l t o f t h e f o u r t h o r d e r

    m e t h o d i s t a k e n a s i n i ti a l g u e s s f o r t h e n e x t s te p . T h e d i f f e r e n t o r d e r s o f t h e t w o

    f o r m u l a s l e a d t o a n e s t i m a t i o n o f t h e lo c a l t r u n c a t i o n e r r o r E S T o f t h e t h i r d

    o r d e r m e t h o d , i n a n a l o g y t o [ 6 ] a n d [ 7 ] . U s i n g t h i s i n f o r m a t i o n t h e f o l l o w i n g

    s t e p s i z e c o n t r o l f o r a g i v e n t o l e r a n c e T O L i s p r o p o s e d .

    hne w:= 0. 9 hold ( T O L ~i- (2 .9)

    \ E S T /

    if hnew g r e a t e r 1 .5 h ola then hnew. '= 1.5 ho~d

    if h .e w le ss 0.5 ho~d t h e n h ,~w :=0 .5 ho~d.

  • 7/25/2019 Generalized Runge Kutta Method

    4/14

    58 P. Kap s and P. Rentrop

    T h e s a f e t y f a c t o r 0 . 9 s e r v e s t o k e e p h n ew s m a l l e n o u g h t o b e a c c e p t e d , i f t h e

    t r u n c a t i o n e r r o r i n th e n e x t s t ep i s g r o w i n g . T h e b o u n d s 0 .5 a n d 1.5 f o r t h e r a t i o

    o f t w o s t e p s a r e i n t r o d u c e d t o p r e v e n t a s te p s i z e p r e d i c t i o n , w h i c h i s h i g h l y z ig -

    z a g i n c h a r a c t e r . T h e v a l u e s o f t h e t h r e e c o n s t a n t s a r e f i xe d b y e x p e r i e n c e . E S T

    i s d e f i n e d b y

    n

    E S T : = m x

    [ Y i h ( X ~ - - f ; i h ( X ~

    i = 1 S i

    w h e r e S = ( S 1 . . . . S ,,) T s t a n d s f o r a s u i t a b l e s c a l i n g v e c t o r .

    S i : = m ax C , ly /,h X/)[ ) i = 1 , . . . , n 2 .10)

    C > 0 in t he f ol lo w i ng C = I )

    x o < x j < Xo la , x j r e p r e s e n t s t h e d i s c r e t e a b s c i s s a .

    h , e w is a c c e p t e d , i f E S T < T O L , o t h e r w i s e f o r m u l a 2 .9 ) is a p p l i e d o n c e m o r e

    w i t h t h e v a l u e E S T b e l o n g i n g t o t h e f a i l e d s t e p s i z e p r e d i c t i o n . T h i s y i e l d s a

    s m a l l e r h . . . . w h i c h m a y b e s u c ce s sf u l. R e p e a t e d a p p l i c a t i o n o f 2 .9) is t e r m i n a -

    t ed , i f hnew=

  • 7/25/2019 Generalized Runge Kutta Method

    5/14

    Generalized

    o r d e r 1 :

    o r d e r 2 :

    o r d e r 3 :

    o r d e r 4 :

    o r d e r 5 :

    Runge Kutta Methods

    Z c i = l ,

    ~ . c i f li = 8 9 Y = P : ( Y ) ,

    Z Ci f l i j f l j _____1 _ 7 + 72 = P 4 ( 7 ) ,

    - - g - - ~ 7 - - - - P 6 (7 ),

    Z Ci f l ik 0~2 1 1

    p

    = ~ - ~ 7 = 7(7),

    ( fo r t r u n c a t i o n e r r o r i n v e s t i g a t io n s )

    C O~4 - 1

    i i - - 5

    Z c l a~ O~,k l k = ~ - - 88 = P1 o(Y ) ,

    2 C i O ~ i k f l k O ~ i l f l l = 2 ~

    1 1 2

    ~ 7 + ~ - 7 = P l y ( Y ) ,

    2 1

    Z C i O~iO~ikO~k - - T 5 ,

    Z C i Xi O~ik lk l f l l 1 1. - - 1 .2

    Z c i f l i k c ~ 3 =T6 7 -1 8 8 = p 1 4 (7 ) ,

    f l l - - T 6 - - ~ 7 + ~ 7

    = P 1 5 ( 7 ) ,

    ~ = ~ - - g r ' - 5 ~ = P ~ 6( 7 ) ,

    59

    3 . 1 )

    3 . 2 )

    3 . 3 )

    3 . 4 )

    3 . 5 )

    3 . 6 )

    3 . 7 )

    3 . 8 )

    3 . 9 )

    3 . 1 0 )

    3 . 1 1 )

    3 . 1 2 )

    3 . 1 3 )

    3 . 1 4 )

    3 . 1 5 )

    3 . 1 6 )

    3 . 1 7 )

    . C , f l , k f l k , f l , m B m - - 1 . - - 2

    - l ~ g - - g ~, -t- y - - 2 7 3 + 7 4 = P 1 7 ( 7 )

    s u m m a t i o n i n d i c e s i, j , k , l , m = 1 . . . , s ,

    a b b r e v i a t i o n s

    f l i j = a l j + 71 j , ~ = 71 i = 0 f o r i < j .

    R e m a r k . T h e o r d e r c o n d i t i o n s f o r Yh a r e o b t a i n e d b y r e p l a c i n g c i b y c i a n d s b y g

    in t h e a b o v e e q u a t i o n s . T h e y a r e d e n o t e d b y (? ). F r o m ( 3.4 ) a n d ( 3.8 ) i t f o l lo w s

    i m m e d i a t e ly , t h a t t h e r e a r e n o m e t h o d s o f o r d e r 4 w i th s = 2 . A - s t a b l e m e t h o d s

    o f o r d e r 4 e x i s t fo r s = 3 , s e e [ 1 3 , 1 4] . T h e r e is n o e m b e d d e d m e t h o d (2 .8 ) w i t h

    s = 3 , ~_< 3 , s e e L e m m a ( 3 .t 8 ) . N e v e r t h e l e s s , o n e c a n c o n s t r u c t m e t h o d s w i t h s = 4,

    g = 3 a n d o n l y t h r e e f u n c t i o n e v a l u a t i o n s p e r s te p , se e P r o p o s i t i o n s ( 3.1 9) , (3 .2 0) .

    I n [1 3 , 1 4 ] i t i s s h o w n t h a t t h e r e e x i s ts n o f i v e o r d e r m e t h o d ( 2.2 ) w i t h s - - 4 .

    Lemma ( 3 .1 8 ). T h e r e e x i s ts n o e m b e d d e d m e t h o d ( 2. 8) w i t h s = 3, ~_< 3 .

    P r o o f

    g = 2 i s i m p o s s i b l e , b e c a u s e t h e z e r o s o f P 4 (7 ) a n d P s (7 ) a re d i f fe r e n t.

    L e t ~ = 3 , t h e n ( 3 . 1 ) , ( 3 . 3 ) a n d ( 3 . 4 ) f o r

    Y h

    a n d Yh d e f in e t h e s a m e l i n e a r s y s t e m

    f o r t h e q , r e s p . cl ( i = 1 , 2 ,3 ) . D u e t o ( 3 .7 ) a n d ( 3. 8) t h e s y s t e m h a s a u n i q u e

    s o lu t io n c ~ = ~ . 9

    P r o p o s i t io n ( 3.1 9) . T h e r e e x i st e m b e d d e d m e t h o d s ( 2.8 ) w i t h s = 4 , g = 3 a n d t h r e e

    f u n c t i o n e v a l u a t i o n s p e r s t e p . T h e p a r a m e t e r s 7 , ~ 2 , a a , c 4 a n d fl4 3 a r e f r e e

    ( e x ce p t f o r s p e c i a l v a l u e s l e a d i n g t o n o n s o l v a b l e l i n e a r s y s te m s ) .

    P r o o f

    C h o o s i n g ~ 4 = ~ 3 , w i t h a , ,1 = c % 1 , a 4 2 = a 3 2 , c q~ 3 = 0 , o n e g e t s a fo u r s t a g e

    m e t h o d w i t h o n l y t h r e e f u n c t i o n e v a l u a t i o n s . E q u a t i o n s (3 .1 ), (3 .3 ), (3 .5 ) d e t e r m i -

  • 7/25/2019 Generalized Runge Kutta Method

    6/14

    60 P . Kap s and P . Ren t rop

    n e C l ,

    C2

    C 3 J - C 4 :

    ~ ~

    c2 =

    -~.

    ~ a~/ c 3 + c 4 88

    F r o m (3 .8 ) f o l l o w s : 1 33 2/3 2= P s ( Y ) / c 4 / 3 4 3 = u . E q u a t i o n s ( ~ 1 ) , ( 3. 3), (3 7 4) d e f i n e C a ,

    c2 , Ca :

    0 c 3 / 4 ( 7 )

    ( 3. 7) l e a ds t o c 3 /332 + c 4 /342 = ( P 7 ( 7 ) - c 4 /343 a 2)/c ~2 = : v . F r o m ( 3.4 ) a n d (3 . 2 ) on e

    o b t a i n s / 3 2 a n d / 3 3 :

    v c / 3 3 1

    e 2 , 3 ,

    t h e r e f o r e /3 3 2 , / 3 4 2 a r e g i v e n b y :

    % 2 c a n b e c a l c u l a t e d f r o m ( 3.6 ):

    ~ 3 2 = P 6 ( 7 ) / ( ( c 3 + c 4 ) ~ 3 / 3 2 ) . / 3 4

    f o l l o w s f r o m

    ( 3 . 2 ) : / 3 4 = ( P 2 ( 7 ) - c 2 / 3 2 - c 3 / 3 3 ) / c 4 9 #

    T h e r e m a i n i n g f re e p a r a m e t e r s , e x c e p t 7 , c a n b e c h o s e n s o t h a t s e v e ra l e q u a t i o n s

    o f c o n d i t i o n o f o r d e r f iv e a r e s a t is f ie d . T h e c o e f f ic i e n t 7 d e t e r m i n e s e s s e n t i a ll y

    t h e s t a b i l i t y p r o p e r t i e s , s e e (2 .5 ), (2 .6 ). c 4 h a s n o i n f l u e n c e o n t h e t r u n c a t i o n

    e r r o r , i ts v a l u e c a n b e c h o s e n t o c 3 = 0 .

    P r o p o s i t i o n ( 3 .2 0 ). T h e f re e p a r a m e t e r s o f P r o p o s i t i o n ( 3 .1 9 ) a r e c h o s e n a s :

    ~2 = 2 y,

    51 I Z 2

    ~ 3 - - 1 1

    4 3-~2

    c 4 a n d / 3 4 3 a r e s o l u t i o n s o f th e l i n e a r s y s t e m :

    T h e n E q s . (3 .9 ), (3 . 1 0 ), (3 .1 1 ), ( 3 .1 4 ) a n d ( 3 . 1 5 ) a r e s a t i s f i e d , t o o . I t h o l d s c 3 = 0 .

    P r o o f

    F o r a 2 : # ~ 3 t h e E q s . (3 .1 ), (3 .3 ), ( 3.5 ) a n d ( 3 .9 ) p o s s e s s a s o l u t i o n , i ff :

    : : I i l

    e t a 2 a ~

    = 0 .

  • 7/25/2019 Generalized Runge Kutta Method

    7/14

    Generalized R unge-K uttaMethods 61

    T h u s t h e c h o i c e o f % i m p l ie s (3 .9 ). D u e t o t h e l in e a r s y s t e m i t h o l d s c 3 = 0 a n d

    (3 .14). T h e c h o i c e o f % c o m bi n e d w i t h ( 3 .6 ) l e a ds t o t he s i m p l i f y i ng a s s um p-

    t i ons :

    k l

    Th ere for e (3 .9) im pl ies (3 . t0) , (3 .11). (3 .14) impl ies (3 .15) , see a l so [13 ] , Pr op os i -

    t io n 5, p. 20. 9

    R e s t r i c t i o n . E v a l u a t i o n o f th e r i g h t h a n d s id e o f t h e d i f fe r e n ti a l e q u a t i o n m e r e l y

    i n t he i n t e g r a t i on i n t e r va l r e qu i r e s :

    0= 0 a nd b > 0,

    s t a b l e a t i n f in i t y , iff: a > 0

    ( a n a lo g u e r e s u l t fo r t h e m e t h o d o f o r d e r 3 w i t h & /~ ). C o m p u t a t i o n o f th e z e r o s

    of a a nd b , r e sp . f i an d ~ l eads t o t he s t ab i l i t y i n t e rv a l s o f ~, l i s ted i n T able (3.24).

  • 7/25/2019 Generalized Runge Kutta Method

    8/14

    62 P. K aps and P. Ren trop

    T a b l e 3 . 2 4 ) . Stabil i ty Intervals

    A -stability Stab ility at infinity,

    see als o sk etch (3.25)

    p = 3 [ 89 [0.15332,~] u [89 oo[

    p = 4 [ 0 . 3 9 4 3 4 ,1 .2 80 57 ] [0 .10 56 7 , .10727]

    u[0.20385 , 0.25] ~[0.3943 4, oo [

    T h e i n t e r v a l s o f A - s ta b i l it y c o r r e s p o n d t o th e r e c e n t l y p u b l i s h e d v a l u e s o f

    B u r r a g e [ 4 ] , T a b l e 1 .

    S k e t c h ( 3 , 2 5 ) . L a g u e r r e P o l y n o m i a l s L 3 ( ~ ) , L 4 ( ~ )

    14.00

    12.00

    10.00

    8.00

    6.00

    4.00

    2.00

    0.00

    2.00

    4.00

    6.00

    8 .00

    10.00

    12. O0

    14.00

    _

    L J 1 / ~

    i / 1_~(1/~,)

    F o r t h e c h o i c e o f ? , b e s i d e s t a b il i t y c o n s i d e r a t i o n s t h e t r u n c a t i o n e r r o r w a s

    t a k e n i n to a c c o u n t . F o r a m e t h o d o f o r d e r p th e l o c a l t r u n c a t i o n e r r o r i s g iv e n

    b y :

    Np l

    hp+I ~ TiP+l~zf+lDf+l+O hp+2),

    s e e [ 3 ] .

    i =

  • 7/25/2019 Generalized Runge Kutta Method

    9/14

    G e n e r a l i z e d R u n g e - K u t t a M e t h o d s 6 3

    T he num e r i c a l c ons t a n t s T ~p+ ~ a r e d e t e r m i n e d b y t h e p a r a m e t e r s o f t h e m e t h o d ,

    D~+ ~ a r e t h e e l e m e n t a r y d i f fe r e n ti a ls o f o r d e r p + 1, e ~ + ~ t h e c o r r e s p o n d i n g

    c oe f f i c i e n t s o f B u t c he r a nd

    Nv ~

    t h e n u m b e r o f tr ee s o f o r d e r p + l . T h e

    f o l low i ng e xp r e s s i on de f i ne s t he e r r o r c on s t a n t :

    6 = m ax [T/p+ 11 (3.2 6)

    T w o m e t h od s w i t h d i f f e r e n t 7 a n d d i f f e re n t s t a b i l it y p r ope r t i e s a r e p r op os e d . I f

    b o t h R O W - m e t h o d s ( 2 .8 ) s h o u l d b e A - s ta b le , s m a l l v a lu e s o f 7 l e a d t o s m a l l

    t r u n c a t i o n e r r o rs . T h e r e f o r e 7 = 0 . 3 9 5 is p r o p o s e d . T h e h y p o t h e s e s o f P r o p o s i -

    t i on (3.20) g ive e3 < 0 , co n t r ad i c t i n g re s t r i c t i on (3.21). In Ta ble (3.27) a coe f f i c i en t

    set fo r 7 = 0 .395 w i th s m a l l t ru nc a t io n e r rors is l is t ed , which do n t s a t i s fy (3.20).

    T a b l e (3 .2 7) . G R K 4 A , ~ = 0 . 3 9 5

    = 0 . 3 9 5 ~ 21 = - 0 . 7 6 7 6 7 2 3 9 5 4 8 4

    Y31 = - 0 . 8 5 1 6 7 5 3 2 3 7 4 2 ~ 32 = 0 . 5 2 2 9 6 7 2 8 9 1 8 8

    7 41 = 0 . 2 8 8 4 6 3 1 0 9 5 4 5 ~ 42 = 0 . 8 8 0 2 1 4 2 7 3 3 8 1 E - 1

    ~43 = - 0 . 3 3 7 3 8 9 8 4 0 6 2 7

    ~ 2 1 = 0 . 4 3 8

    ~ 3 1 = 0 .7 9 6 92 0 4 57 9 3 8 e 32 = 0 . 7 3 0 7 9 5 4 2 0 6 1 5 E - 1

    c l = 0 . 3 4 6 3 2 5 8 3 3 7 5 8 c 2 = 0 . 2 8 5 6 9 3 1 7 5 7 1 2

    c3 = 0 . 3 6 7 9 8 0 9 9 0 5 3 0

    c 1 = 0 . 1 9 9 2 9 3 2 7 5 7 0 1 c 2 = 0 . 4 8 2 6 4 5 2 3 5 6 7 4

    c a = 0 . 6 8 0 6 1 4 8 8 6 2 5 6 E - 1 c 4 = 0 .2 5

    6 N 0 . 9 4 2 / 5 f o r t h e m e t h o d o f o r d e r 4 , s e e ( 3 .2 6 )

    S N 1 . 0 8 /4 f o r t h e m e t h o d o f o r d e r 3 , s e e ( 3 .2 6 )

    T h e s e c o n d m e t h o d i s c o n s t r u c t e d a c c o r d i n g t o P r o p o s i t i o n ( 3 . 2 0 ) .

    7 e [ 0 .1 0 5 6 7 ,0 . 1 0 7 2 7] p r o d u c e g r e a t v a l u e s o f L 3 ( 1 ) . F o r

    7e [0 .20385 , 0 .25]

    L 3 ( ~ ) i s s m a l t , s e e s k e t c h ( 3 . 2 5 ) , a n d t h e s t a b i l i t y r e g i o n o f t h e f o u r t h o r d e r

    m e t h o d i s v e r y l a r g e . F o r 7 = 0 . 2 3 1 t h e f o u r t h o r d e r m e t h o d i s A ( 8 9 . 3 ~

    and the h yp oth ese s o f (3 .20) and re s t r i c t i on (3.21) a re sa t i s fi ed . A co e f f i c i en t se t is

    l is t ed i n Tab le (3.28). A fur th e r re l a t ed co e f f i c i en t se t wi th 7=0 .22 042 841 can b e

    found in Stoer , Bul i rsch 1-19] .

    ab l e

    (3 .2 8) . G R K 4 T , ~ = 0 . 2 3 1

    7 = 0 .2 3 1 Y21 = - 0 . 2 7 0 6 2 9 6 6 7 7 5 2

    ~ al = 0 . 3 1 1 2 5 44 8 3 2 9 4 T32 = 0 . 8 5 2 4 4 5 6 2 8 4 8 2 E - 2

    741 = 0 .2 8 2 8 1 6 8 3 2 0 4 4 Y 42 = - 0 . 4 5 7 9 5 9 4 8 3 2 8 1

    743 = - 0 . 1 1 1 2 0 8 3 3 3 3 3 3

    % ~ = 0 . 4 6 2

    % 1 = - 0 . 8 1 5 6 6 8 1 6 8 3 2 7 E - 1 ~ a2 = 0 .9 6 1 77 5 1 50 1 6 6

    6~ = - 0 . 7 1 7 0 8 8 5 0 4 4 9 9 62 = 0 . 1 7 7 6 1 7 9 1 2 1 7 6 E + 1

    ~3 = - 0 . 5 9 0 9 0 6 1 7 2 6 1 7 E - 1

    c~ = 0 . 2 1 7 4 8 7 3 7 1 6 5 3 c 2 = 0 . 4 8 6 2 2 9 0 3 7 9 9 0

    c a = 0 . c 4 = 0 . 2 9 6 2 8 3 5 9 0 3 5 7

    6 ~ 0 . 1 9 9 / 5 f o r t h e m e t h o d o f o r d e r 4 , s e e ( 3. 26 )

    8 ~ 0 . 4 6 1 / 4 f o r t h e m e t h o d o f o r d e r 3 , s e e ( 3. 26 )

  • 7/25/2019 Generalized Runge Kutta Method

    10/14

    6 4 P . K a p s a n d P . R e n t ro p

    4 Nu m erical Implementation

    T o c o m p u t e t h e v e c t o r s k i 2 .2 ), a li n e a r s y s t e m o f o r d e r n fo r f o u r r ig h t h a n d

    s i d e s m u s t b e s o l v e d . I n o r d e r t o a v o i d m a t r i x - v e c t o r m u l t i p l i c a t i o n s , t h e

    e q u i v a l e n t f o r m d u e t o [ 2 1 ] i s u s e d :

    ( I -- h Y f ( Y o ) ) k , = h f ( Y o ) ,

    ( I - h y f ( y o ) ) ( k 2 + ~ 2 , k l ) = h f ( Y o + a21 k 1) ~- ~21 k , ,

    ( I - h y f ( y o ) ) (k 3 +

    ~3~ k~ + ~32 k2 ))

    = h f ( Y o + C t 3 1 k ~

    + a 32 k2 ) + ~ 3 1k l + Y 32k2),

    ( I - h v f ( y o ) ) ( k 4 + 74~ k~ + ~42 k2 + ~43 k3))

    = h f ( Y o + a 3 a k ~ - ] - a 3 2

    k2 ) -b (~41 k 1 1- ~42 k2 -I - ~43 k3 )

    w h e r e

    T h e J a c o b i a n f ( Y o ) is c o m p u t e d b y d i ff e re n c e a p p r o x i m a t i o n a n d s h o u l d b e

    r e p l a c e d b y a n a n a l y t i c v e r s i o n f o r v e r y s e n s i t i v e p r o b l e m s . T h e m a t r i x

    1 - h

    7 f ( Y o ) )

    is d e c o m p o s e d b y L U - f a c t o r iz a t io n . C o m p u t a t i o n o f t h e k is e q u iv a -

    l e n t to b a c k s u b s t i tu t i o n s . F o r l a r g e s p a r s e s y s t e m s th e s t r u c t u r e o f t h e J a c o b i a n

    i s s a v e d a n d t h e s t a n d a r d r o u t i n e f o r L U - d e c o m p o s i t i o n s h o u l d b e e x c h a n g e d

    b y s u b ro u t in e s fo r s p ar se s ys te m s . B o t h p r o g r am s G R K 4 A 3.27) a n d G R K 4 T

    3 . 28 ) ha ve a s t r uc t u r e a s s i m p l e a s t he R K F m e t hods [ 6 , 7 ] a nd c a n be e a s i l y

    i m p l e m e n t e d . E x c e p t f o r g e n e r a t i o n o f t h e J a c o b i a n n o n e s t e d lo o p s a r e n e ce s-

    s a r y . T he c a l l i ng s e que nc e i s i n a c c o r da nc e w i t h [ 6 , 7 ] .

    5 Test Exam ples

    T h e p r o p o s e d m e t h o d s w e r e t e s t ed o n 2 5 st if f d i f fe r e n ti a l e q u a t i o n s [ 8 ]. T h e

    p r op e r t i e s o f t he d i f f e r e n t ia l e qu a t i on s a r e on l y b r ie f l y de s c r i be d i n t he f o l low -

    i ng , f u r t he r i n f o r m a t i ons c a n be f ound i n [ 8 ] . T he t e s t s e t i s d i v i de d i n t o f i ve

    c lasses :

    C l a s s A : L i ne a r w i t h r e a l e i ge nva l ue s

    C l a s s B : L i n e a r w i t h n o n - r e a l e i g e n v a lu e s .

    C l a s s C : N o n l i n e a r c o u p l i n g w i t h r e a l e i g en v a t u es .

    C l a s s D : N on l i ne a r w i t h r e a l e i ge nva l ue s .

    C l a s s E : N o n l i n e a r w i t h n o n - r e a l e i g e n v a l u e s .

    T h e f o l lo w i n g a b b r e v i a t io n s a r e u s e d :

    T Z : T o t a l c o m p u t i n g t i m e i n s e c o n d s t o s o l v e a p r o b l e m . C o m p u t a t i o n s

    w e r e p e rf o r m e d i n F O R T R A N s in g le p r e c is io n w i th a 38 b i t m a n ti ss a

    11 d e c im a l s ) o n t h e T R 4 4 0 o f t h e L e i b n i z R e c h e n z e n t r u m d e r B a ye ri-

    s c h e n A k a d e m i e d e r W i s s e n s c h a f t e n .

  • 7/25/2019 Generalized Runge Kutta Method

    11/14

    G e n e r a l i z ed R u n g e - K u n a M e t h o d s 6 5

    F CN N u m b e r o f f u n c ti o n c a ll s

    FJAC N u m b e r o f J a c o b i a n e v a lu a ti o n s. O n e e v a l u a t io n o f t h e J a c o b i a n c o s t s

    n fu n c t io n ca l l s .

    TF

    T o t a l n u m b e r o f f u n c t i o n c a ll s:

    T F = F C N + n - FJ A C .

    LU N u m b e r o f L U - d e c o m p o s i t i o n s , e q u iv a l e n t to t h e n u m b e r o f s te ps .

    ERR M a x i m u m e r r o r o f s o l u ti o n c o m p o n e n t s a t t h e e n d o f t h e in t er v al . T h e

    r ef e re n c e s o lu t io n w a s c o m p u t e d b y t h e p r o c e d u r e D R I V E w i th T O L

    = 1 . E - 8 . D R I V E i s t h e i m p r o v e d G E A R v e r s i o n f r o m 1 3 . 1 . 1 9 7 5 , d u e t o

    G e a r 1 -9 ] a n d H i n d m a r s h [ 1 2 ] .

    F o r a l l e x a m p l e s t h e i n i t i a l s t e p s i z e H I = 1 . E - 3 a n d t he st ep s iz e c o n t r o l

    fo rm u la 2 .9 ) we re u sed . Th e t e s t r e su l t s a r e l i s t ed in Tab le 5 .1 ) an d Tab le 5 .2 ) .

    B o t h m e t h o d s s o l v e a l l e x a m p l e s r e l i a b l y . G R K 4 A l o o s e s p r e c i s i o n i n D 5

    a n d E 2 . A c c o r d i n g t o p re c i s i o n a n d f a s tn e ss , G R K 4 T is t h e s u p e r i o r m e t h o d , in

    s p i te o f i ts w e a k e r s t a b i li ty c o n d i t io n s . O n l y i n E 4 , th e c o m p u t i n g t i m e o f

    G R K 4 T is e n la r ge d . A n o v e r a ll s u m m a r y f o r b o t h m e t h o d s a n d f o r t h r e e

    t o l e r a n c e s i n a c c o r d a n c e w i t h [ -8 ] i s g i v e n i n T a b l e 5 .3 ).

    T a b l e 5 .1 ). S t a t i s t i c s f o r e a c h p r o b l e m , T O L = 1, E - 4

    P r o b l e m G R K 4 A

    T Z L U F C N F J A C T F E R R

    A I 0 . 3 1 4 0 1 2 0 4 0 2 8 0 2 .1 E - 7

    A 2 1 .2 2 5 3 1 55 4 9 5 9 6 4 . 8 E - 8

    A 3 0 . 4 4 6 0 1 7 5 5 5 3 9 5 1 .1 E - 6

    A 4 2 .0 1 7 4 2 1 3 6 5 8 6 3 1 . 5 E - 6

    B 1 1 . 4 4 1 8 3 5 4 2 1 7 6 1 , 2 4 6 4 . 0 E - 5

    B 2 0 . 4 9 4 0 1 2 0 4 0 3 6 0 1 . 0 E - 6

    B 3 0 .5 3 4 3 1 2 9 4 3 3 8 7 7 . 5 E - 7

    B 4 0 . 77 6 2 1 86 6 2 5 5 8 1 . 1 E - 6

    B 5 2 . 0 4 1 6 4 4 9 2 1 6 4 1 , 4 7 6 2 . 4 E - 6

    C 1 0 . 3 6 4 5 1 3 5 4 5 3 1 5 1 . 7 E - 7

    C 2 0 . 37 4 3 1 2 9 4 3 3 01 3 . 6 E - 7

    C 3 0 . 4 5 5 3 1 5 9 5 3 3 7 1 1 .1 E - 5

    C 4 1 . 0 4 1 2 2 3 6 6 1 2 2 8 5 4 9 . 7 E - 6

    C 5 1 . 3 2 1 5 4 4 6 2 1 5 4 2 , 9 1 9 1 . 2 E - 5

    O 1 1 . 1 9 2 0 7 6 2 1 2 0 7 1 , 2 4 2 5 . 9 E - 6

    D 2 0 . 4 6 7 8 2 3 1 7 5 4 5 6 7 . 2 E - 5

    D 3 0 . 3 8 4 9 1 4 0 4 2 3 0 8 1 . 5 E - 7

    D 4 0 . 1 4 2 5 7 5 2 5 1 5 0 1 . 8 E - 5

    D 5 0 .1 1 2 8 8 4 2 8 1 4 0 8 . 7 E - 3

    D 6 0 . 13 2 3 6 9 2 3 1 3 8 1 . 8 E - 4

    E 1 1 .8 1 2 8 8 5 5 2 1 8 4 1 , 28 8 3 . 4 E - 1 0

    E 2 0 . 33 8 7 2 5 0 7 6 4 0 2 8 . 9 E - 4

    E 3 0 . 4 8 8 0 2 3 8 7 8 4 7 2 4 . 7 E - 5

    E 4 0 . 79 7 8 2 2 9 7 3 5 2 1 3 . 5 E - 4

    E 5 0 . 2 6 3 3 9 9 3 3 2 3 1 3 . 0 E - 8

  • 7/25/2019 Generalized Runge Kutta Method

    12/14

    Table 5.2). Statistics for each probem, TO L= 1. E- 4

    P. Kaps and P. Rentrop

    Problem GRK4T

    TZ LU FCN FJAC TF ERR

    A1 0.25 35 105 35 245 1.6E-6

    A2 1.14 51 148 46 562 2 .1E-7

    A3 0.39 54 157 49 353 2 .0E- 5

    A4 1.76 65 186 56 756 2.0E - 5

    B 1 1.24 160 473 153 1,085 3.6 E - 5

    B2 0.45 36 108 36 324 1.2E-6

    B 3 0.46 38 114 38 342 1.6E- 6

    B 4 0.66 53 159 53 477 2.0 E - 6

    B 5 1.75 140 420 140 1,260 1.7 E - 6

    C1 0.31 41 123 41 287 1.8E-7

    C2 0.32 39 117 39 273 5 .0E- 8

    C 3 0.44 53 159 53 371 1.5 E - 7

    C4 0.94 111 333 111 777 1.2E - 7

    C5 1.19 135 405 135 945 4 .5E- 8

    D 1 1.29 231 658 196 1,246 3 .8E- 6

    D2 0.36 63 182 56 350 4 .9E-5

    D3 0.45 57 164 50 364 3 .2E- 8

    D4 0.14 25 75 25 150 2 .2E-6

    D5 0.14 36 104 32 168 1.1E-4

    D6 0.10 17 51 17 102 2 .9E-6

    E 1 1.08 168 327 109 763 3 .4E- 10

    E 2 0.35 96 268 76 420 4.6 E - 4

    E 3 0.49 89 249 71 462 4.3 E - 6

    E4 3.27 354 942 234 1,878 9 .7E- 5

    E5 0.27 33 99 33 231 3 .3E-8

    Table 5.3). Overall Summary

    Method TOL TZ LU FCN FJAC TF

    GRK 4 A 1.E - 2 9.39 1,066 2,851 927 6,859

    1.E-4 18.85 2,112 5,971 1,955 14,428

    1 .E 6 65.35

    7,271 21,189 7,088 49,324

    GRK4T 1.E - 2 8.40 960 2,666 864 6,419

    I . E - 4 19.23 2,180 6,126 1,884 14,181

    1.E - 6 58.07 6,676 19,179 6,320 43,336

    B o t h m e t h o d s a r e l o w o rd e r m e t h o d s a n d w o rk v e ry w el l f o r l o w t o le r a n ce s .

    Th e y s h o u l d b e u s e d o n l y for t o l e r a n c e s b e t w e e n 1 .E -2 a n d 1 .E -4 .

    Com par i so n wi th DRI VE [9, 12] :

    Th e i m p ro v e d G EA R v e r s i o n D R IV E is a v a i l a b l e t o t h e a u t h o r s . C o m p u t i n g

    t ime and func t ion ca l ls fo r a ll exa mple s a re l i s t ed in Tab le 5.4).

  • 7/25/2019 Generalized Runge Kutta Method

    13/14

    G e n e r a l i z e d R u n g e - K u n a M e t h o d s

    able

    5 .4 ). O v e r a l l S u m m a r y f o r T O L = 1 .E - 4

    G R K 4 A G R K 4 T D R I V E - G E A R

    Al l ex am ple s T Z 18 .85 19 .23 44 .21

    T F 14,428 14,181 9,099

    Al l ex am ple s T Z 16 .02 14 .21 21 .57

    exc ept B5, E4 T F 12,43 t 11,043 5 ,423

    67

    Comparing computing time, both methods are competitive with DRIVE,

    although the number of function calls TF is enlarged by a factor two. DRIVE

    runs very efficient in the classes D and E, but produces heavy difficulties in B5,

    where precision is lost and computing time reaches 21.23 seconds. The great

    number of evaluations of the Jacobian and LU-decompositions in GRK 4 A and

    GRK4T are disadvantageous for large complicated systems, which are not

    included in the test set [8].

    6 Applicat ion o f G RK 4A and G RK 4T

    to the restricted Three Body Problem

    To give some information how both methods will work for non-stiff differential

    equations, the restricted Three Body Problem earth-moon-spaceship) tested in

    Bulirsch, Stoer [3] and [7] was solved. Results for the non-stiff differential

    equation solvers DIFSY1, VOAS, RKF7, and RKF4 from [6] together with

    GR K4 A and GRK4T are listed in Table 6.1).

    able

    6.1). T h r ee B o d y P r o b l em , T O L = 1 . E - 4 ,

    HI

    1 . E - - 3

    S ta tis ti c s D I F S Y 1 V O A S R K F 7 R K F 4 G R K 4 A G R K 4 T

    TZ 1.11 2.36 1.28 1.44 1.89 2.48

    TF 1,215 669 1,233 1,398 1,048 1,339

    This difficult example, which requires a robust and reliable stepsize controll,

    was solved precisely by both methods. The more complicated structure of

    GRK4A and GRK4T enlarged the computing time by a factor 1.5, although

    the number of function calls is comparable to the related routine RKF 4.

    Conclusion

    With GRK4A and GRK4T two reliable, fast and precise algorithms for the

    numerical solution of stiff systems of ordinary differential equations are avail-

    able. The loworder methods should be applied for low tolerances up to TOL

    t .E-4 . One would prefer these methods for problems with n < 10, because of

  • 7/25/2019 Generalized Runge Kutta Method

    14/14

    6 8 P . Ka p s a n d P . Re n t r o p

    t h e la r g e n u m b e r o f L U d e c o m p o s i t i o n s . I f s t a b il it y r e q u i r e m e n t s a re w e a k e r

    G R K 4 T s e e m s t o b e t h e f as te r a n d m o r e p r e c is e r o u t in e .

    Acknowledgement

    T h e a u th o r s w i s h t o t h a n k P r o f . R . Bu l i r s c h . Dr . E . Ha i r e r a n d P r o f . G . W a n n e r

    fo r insp i r ing and he lp fu l d i scuss ions .

    R e f e r e n c e s

    1. Ab r a m o w i t z , M ., S te g u n , I.A . : H a n d b o o k o f m a th e m a t i c a l f u n c ti o n s . Ne w Yo r k : D o v e r P u bi .

    Inc. 1970

    2 . Be d e t , R .A ., E n r ig h t , W .H . , Hu l l , T .E .: S T I F F DE T E S T : A p r o g r a m f o r c o m p a r in g n u m e r i c a l

    m e th o d s f o r s t if f o r d in a r y d i f f e re n t i a l e q u a t i o n s . T e c h . Re p . N o . 8 1 , Un iv e r s i t y o f T o r o n to 1 97 5

    3 . Bu l i r s c h , R . , S to e r , J . : Nu m e r i c a l t r e a tm e n t o f o r d in a r y d i f f e r e n t i a l e q u a t i o n s b y e x t r a p o l a t i o n

    me thods . Num er . Ma th . 8 , 1 -13 1966)

    4 . Bu r r a g e , K . : A s p e c ia l f a m i ly o f R u n g e - K u t t a m e th o d s f o r s o lv in g s t if f d i f fe r e n t ia l eq u a t io n s .

    B IT 18, 22-4 1 1978)

    5 . Da h lq u i s t , G . : A s p e c ia l p r o p e r ty f o r l i n e a r m u l t i s t e p m e th o d s . B1 T 3 , 2 7 - 4 3 1 96 3)

    6. Diek hoff , H.-J . , Lory, P . , O berle , H.J ., P esch, H.-J. , R ent rop , P . , Seydel , R.: Co m pa rin g rou tines

    f o r t h e n u m e r i c a l s o lu t i o n o f i n it i a l v a lu e p r o b l e m s o f o r d in a r y d i f fe r e n ti a l e q u a t i o n s i n m u l t i p l e

    shoo t ing . Nu me r . Ma th . 27 , 449-46 9 1977)

    7 . Enr ig h t , W.H., B ede t , R , Fa rk as , I ., Hu l l , T .F . : Tes t re su l t s on in i t ia l va lu e meth ods fo r non-s t i f f

    o r d in a r y d i f f e re n t i a l e q u a t i o n s . T e c h n . Re p . N o . 6 8 , Un iv e r s i t y o f T o r o n to 1 97 4

    8 . E n r ig h t , W .H . , Hu l l , T .E. , L in d b e r g , B . : Co m p a r in g n u m e r i c a l m e th o d s f o r st i ff s y s t e m s o f

    o rd ina ry d i f fe ren t ia l equa t ion s . Tech . Rep . No . 69 , 1974 Univers i ty o f To ron to , see a l so BIT 15 ,

    10-48 1975)

    9 . Ge a r , C .W . : N u m e r i c a l i n i t ia l v a lu e p r o b l e m s i n o r d in a r y d i f fe r e n ti a l e q u a t i o n s . N .Y . : P r e n ti c e

    Hal l 1970

    10. Gr igor ie f f , R .D. : N um er ik gew/Shnl icher Di f fe ren t ia lg le ich ungen . S tu t tga r t : Teu bn er 1972

    11 . Ha i r e r , E ., W a n n e r , G . : O n t h e B u tc h e r g r o u p a n d g e n e r a l m u t iv a lu e m e th o d s . Co m p u t in g 1 3 , 1 -

    15 1974)

    1 2. H in d m a r s h , A .C . : G E A R - o r d in a r y d i f fe r e n t ia l e q u a t i o n s y s t e m s o lve r . UC I D- 3 0 0 0 1 , Re v. 2 ,

    Un iv e r s i t y o f Ca l i f o r n i a : L a wr e n c e L iv e r m o r e L a b o r a to r y 1 97 2

    1 3. Ka p s , P . : M o d i f i z i e r te R o s e n b r o c k m e th o d e n d e r Or d n u n g 4 , 5 u n d 6 z u r n u m e r i s c h e n I n t eg r a -

    t i o n s t e if e r D i f f e r e n t i a lg l e i c h u n g e n . D i s s e r t a t i o n , U n iv er s it~ it I n n s b r u c k , S e p t e m b e r 1 97 7

    14. Kap s , P . , Wa:nner, G. : Ro sen bro ck- t ype me tho ds o f h igh o rder . In p ress 1979)

    1 5. No r s e t t , S .P .: C - p o ly n o m ia l s f or r a t i o n a l a p p r o x im a t io n t o t h e e x p o n e n t i a l f u n c t io n . Nu m e r .

    M ath . 25 , 39 -56 1975)

    1 6. No r s e t t , S .P. , W a n n e r , G . : T h e r e a l - p o l e s a n d w ic h f o r r a t i o n a l a p p r o x im a t io n s a n d o s c i ll a ti o n

    equ a t ions . BIT 19 , 79 -94 1979)

    1 7. No r s e t t , S .P ., W o l f b r a n d t , A . : Or d e r c o n d i t i o n s fo r Ro s e n b r o c k - ty p e m e th o d s . N u m e r . M a th . 32 ,

    1-15 1979)

    18. Ro senb rock , H .H. : Som e gene ra l imp l ic i t p rocesses fo r the num er ica l so lu t ion o f d if fe ren tial

    equa tons . Co mp . J . 5 , 329 331 1963)

    19. S toe r , J. , Bu l i r sch , R. : E in f t ih rung in d ie Num er isc he M ath em at i k I I. Ber l in , He id e lbe rg , New

    Yo rk : Spr in ger Ver lag , 2. Auf lage , 1978

    2 0. W a n n e r , G . : L e t t e r t o S.P. N o r s e t t a n d u n p u b l i s h e d c o m m u n ic a t i o n s t o P . Ka p s a n d A .

    W o l f b r a n d t , 1 9 7 3

    2 1. W o l f b r a n d t , A . : A s tu d y o f Ro s e n b r o c k p r o c e s se s w i th r e s p e c t t o o r d e r c o n d i t i o n s a n d s ti ff

    s tab i l i ty . Dis se r ta t ion , Res ea rch Pep . 77 .01 R, Univ ers i ty o f G/S teborg, M arc h 1977

    Re c e iv e d S e p t e m b e r 2 2 , 1 9 7 8 /Re v i s e d Ap r i l 9 , 1 97 9