Domain- S p c if ic L ang u ag A g nt s Mrik Mrist 1, T õ n is K l d r 1, J ü ri H l kiv i 1, L o Mo tu s 2 1 Univrsity of Tartu, Estonia {Mrik.Mrist, Tonis.Kldr, Jyri.Hlkivi}@ut. 2 Tal l inn Tc h nic al Univrsity, Estonia Lo.Motus@dcc.ttu. Abstract. M ul ti- ag nt systm s ar p rop osing a nw d sig n c onc p t for softw ar th k ntic softw ar d sig n b asd on th c onc p t of intrac tiv ag nts. O n th b asis of th is c onc p tual viw, intrac ting ag nts ap p ar as an ap p rop riat c onc p tual tool for th d vl op m nt of d om ain- sp c ific l ang uag s. D om ainsp c ific l ang uag w il l b c onsid rd as a c onsnsual c ol l c tion of intrrl atd autonom ous notions and, th l ang uag p roc ssor as a c l ustr of intrac ting ag nts i.. a multi- a g n t rp rsnting l ang uag notions. Th is p ap r sug g sts an ag nt- b asd m od l ing fram w ork as a p ossib l m th od ol og ic al b asis for D S L d sig n and d vl op m nt. 1 Introduction A n in trstin g a sp c t in d o m a in - sp c if ic l a n g u a g s ( D S L ) is th a t o f f o rm a l m o d l s a n d f ra m w o rks f o r D S L d sig n a n d f o r th d v l o p m n t o f a p p ro p ria t m o d l s o f l a n g u a g n o tio n s a n d l a n g u a g c o n stru c ts. T o w h a t x tn t th f o rm a l m th o d s a p p l id su p p o rt th ra so n a b l stru c tu rin g o f in f o rm a tio n o b j c ts a n d p ro b l m so l v in g, is a c ru c ia l a sp c t o f m o d l in g. T h f f ic in t im p l m n ta tio n a n d a c c o rd a n c o f th m o d l w ith p ro b l m a ra sp c if ic a tio n l a n g u a g c o n c p ts a r u su a l l y v a l u d m o st in th is p ro c ss. A ra so n a b l f ra m w o rk is x p c td to p ro v id to o l s f o r D S L m o d l in g a n d im p l m n ta tio n o n a c o n c p tu a l l y c l a r b a sis. T h v a rity o f c o n tx tu a l v iw s a n d a p p l ic a tio n a ra s a c c n tu a t th c o m p l ic a td n a tu r o f th ta sk. T h su c c ss o f sy stm m o d l in g d p n d s o n th x tra c tio n o f su rf a c a n d / o r d p ( rg u l a r o r c o n tx tu a l ) su b stru c tu rs f ro m th in f o rm a tio n n v iro n m n t a n d th ir su b sq u n t a tta c h m n t to c o m p u ta tio n a l a c tiv itis. Mu l ti- a g n t sy stm s a r p ro p o sin g a n w d sig n c o n c p t f o r so f tw a r th k n t i c p r o g r a m d s i g n b a sd o n th c o n c p ts o f in tra c tio n s a n d a g n ts.i n th is p a p r w c o n sid r th id a o f kn tic d sig n o f D S L p ro c sso rs. D o - m a in - sp c if ic l a n g u a g is tra td a s a c o l l c tio n o f in trrl a td a u to n o m o u s n o tio n s a n d, its l a n g u a g p ro c sso r a s a c l u str o f in tra c tin g a g n ts i.. a m u l t i - a g n t rp rsn tin g th d o m a in - sp c if ic l a n g u a g b y n o tio n s. O n th b a sis o f th is c o n c p tu a l v iw th in tra c tiv a ttrib u td a u to m a ta c o n c p t is c o n sid rd a s a p o tn tia l p a rtic u l a r f ra m w o rk f o r D S L m o d l in g a n d im p l m n ta - tio n. A ttrib u td a u to m a to n is a sta t tra n sitio n m a c h in in tro d u c d a n d a p p l id in i- tia l l y f o r p u rp o ss o f stru c tu rin g a n d c o n c p tu a l sp c if ic a tio n o f kn o w l d g o n th b a sis o f rg u l a r a ttrib u td stru c tu rs [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 1 0 ]. T o su p p o rt th d c o m p o si-
Domain-S p c if ic L ang u ag A g nt s 8 3 t ion of a s p c if ic at ion or c omp u t at ional t as k int o a n t w or k of c omp on nt s, an at t r ib - u t d au t omat on h as f init m mor y d is t r ib u t d t o it s s t at s, as w ll as c omp u t at ions and c ommu nic at ion ac t ions at t r ans it ions. I nt r ac t iv c omp os it ions of A t t r ib u t d A u t omat a ( A A ) s r v as a t ool w h il t r at ing A A as c omp l x c omp u t at ional ag nt s. B as ic int r ac t iv A A c omp os it ion t c h niq u s ar imp l m nt d in int r ac t iv A A v is u al d v lop m nt nv ir onm nt p r ot ot y p d s c r ib d in t h is p ap r. A t t r ib u t d au t omat on and int r ac t iv at t r ib u t d au t omat on w r lat r on id nt if i d as a k ind of W g n r s s q u nt ial int r ac t ion mac h in and mu lt i-s t r am int r ac t ion mac h in r - s p c t iv ly [ 1 1, 1 2, 1 3 ]. T h int g r at d p r s nt at ion and s p c if ic at ion of k now l d g w ill ob v iou s ly r main a r s ar c h and t c h nolog ic al d v lop m nt p r ob l m, f or DS L t oo. T h u s, on t h on h and, t h d v lop m nt of mod ls f or c onc p t u al, x p r t and p r oc d u r al k now l d g, as w ll as of ap p r op r iat m t h od s f or k now l d g p r s nt at ion and manag m nt is s s n- t ial f or t h s u c c s s f u l u s of d f init DS L in p r ac t ic. O n t h ot h r h and, w w ou ld lik t o f oc u s in t h is c ont x t on t h d omain-s p c if ic lang u ag as an nv ir onm nt f or mod ling and manag ing inf or mat ion ob j c t s and p r ob l m s olv ing ac t ions f or t h p r s - nt at ion of ling u is t ic not ions. T h n x t s c t ion w ill c ons id r d if f r nt v i w s of a d o- main-s p c if ic lang u ag as a p r ob l m-s olv ing nv ir onm nt ( i.., a k ind of c ommu - nit y of p r ac t ic ), a c oll c t ion of int llig nt ag nt s [ 1 4 ], and as a c lu s t r of int r ac t ion mac h in s. S c t ion 3 w ill p r s nt t h A A c onc p t. S c t ion 4 w ill c onc nt r at on t h int r ac t iv A A ap p r oac h as a DS L d v lop m nt and imp l m nt at ion t ool b as d on t h p ar ad ig m of int r ac t iv c omp u t at ions. 2 Agnts, Intractiv Computations and Communitis of Practic D s ig n and imp l m nt at ion m t h od s f or s of t w ar -int ns iv s y s t ms h av u nd r g on a r mar k ab l v olu t ion d u r ing t h las t d c ad s. A lg or it h mic ap p r oac h in c omp u t r s c i nc and s of t w ar ng in r ing h as b n s u b s t it u t d b y int r ac t ion-c nt r d p ar a- d ig ms. I n ar t if ic ial int llig nc a s imilar s h if t, f r om log ic -b as d t o ag nt -b as d mod - ls, h as t ak n p lac. I nt r ac t ion-c nt r d s y s t ms ap p ar t o b mor p ow r f u l p r ob l m s olv r s t h an alg or it h m c nt r d. I t is als o s u g g s t d t h at alg or it h mic mod ls alon d o not s u f f ic t o x p r s s c r t ain x p c t d as p c t s of b h av ior of t od ay s s y s t ms. g. ab ilit y t o s lf -or g aniz t h int r ac t ion of it s c omp on nt s, t o ad ap t it s b h av ior t o t h c h ang s in it s nv ir onm nt, or in it s g oal f u nc t ion. M u lt i-ag nt ap p r oac h or knt i c s [ 1 5 ], is p r op os ing a n w d s ig n c onc p t f or s of t w ar t h knt i c p r o g r a m d s i g n b as d on t h c onc p t of int r ac t ing au t onomou s ag nt s. C onc p t u ally, ac h c omp o- n nt of a s y s t m c an b nv is ag d as a n a g nt w it h it s s k ills and it s g oals and w it h all t h ag nt s at t mp t ing t o r s p ond t h n d s of t h u s r of t h s y s t m. I nt r ac t ing au t onomou s ag nt s as a p ar ad ig m f or s of t w ar d s ig n s u g g s t t h ob v i- ou s id a t o mod l DS L as a c oll c t ion of ag nt s, i.. as a d is t r ib u t d ( ar t if ic ial) int llig nt s y s t m. E ac h c omp on nt of a s y s t m c an b nv is ag d as a n a g nt w it h it s s k ills and it s g oals and w it h all t h ag nt s at t mp t ing t o r s p ond t h n d s of t h u s r of t h s y s t m. T h is int r s t ing ap p r oac h l ad s t o t h p r ob l m of k n t ic d s ig n of a DS L p r oc s s or. W w ill c ons id r a ( d omain-s p c if ic ) lang u ag as a c oll c t ion of int r r - lat d au t onomou s not ions and, t h lang u ag p r oc s s or as a c lu s t r of int r ac t ing p r -
84 M r i k M r i s t, T õ n i s K l d r, J ü r i H l k i v i, L o M o t u s s i s t n t a g n t s i.. a multi- a g n t o f l a n g u a g n o t i o n s. W c o n s i d r t h a t k i n d o f m o d l - i n g f r a m w o r k a s a m t h o d o l o g i c a l l y s u i t a b l b a s i s f o r D S L d s i g n a n d d v l o p m n t t o o l s O n t h o t h r h a n d, a ( d o m a i n - s p c i f i c ) l a n g u a g c r t a i n l y i s a c o n s n s u a l c o l l c - t i o n o f n o t i o n s w h i c h f o r m s t h b a s i s o f t h p r o b l m s o l v i n g n v i r o n m n t f o r p a r t i c u - l a r c o m m u n i t i s o f p r a c t i c. T h c o n c p t o f c o mmun ity o f p r a c tic i s a c n t r a l c o n - c p t o f t h t h o r y o f s o c i a l l a r n i n g [ 1 6 ]. I n t h i s a p p r o a c h, t h p r i m a r y u n i t o f a n a l y s i s i s n i t h r t h i n d i v i d u a l n o r s o c i a l i n s t i t u t i o n b u t t h i n f o r m a l c o m m u n i t y o f p r a c t i c t h a t i s f o r m d f o r s o m a c t i v i t y. T h p a r t i c u l a r p r o b l m - s o l v i n g n v i r o n m n t ( h r D S L ) w i l l s r v a s a h a n d y t o o l t o t h x t n t t h a t i t s i m u l a t s a n d s u p p o r t s t h b a s i c n o t i o n s, p r i n c i p l s a n d x t n s i b i l i t y o f t h l a n g u a g a n d, t h r b y, i n t g r a t s t h k y c o m p o n n t s o f t h s c o m m u n i t i s o f p r a c t i c. I n t r m s o f i n t r a c t i v c o m p u t a t i o n s, t h c o m p u t a t i o n a l a g n t s a n d c o m m u n i t i s o f p r a c t i c a r a k i n d o f i n t r a c t i o n m a c h i n s, a s i n t r o d u c d b y P t r W g n r. T h o u g h i n t r a c t i o n m a c h i n s a r a s i m p l a n d o b v i o u s x t n s i o n o f T u r i n g m a c h i n s, t h i s s m a l l c h a n g i n c r a s s x p r s s i v n s s s o i t b c o m s t o o r i c h f o r m a t h m a t i c a l m o d - l s. I n t r a c t i o n m a c h i n s h a v s i n g l o r m u l t i p l i n t r a c t i o n s t r a m s a n d s y n c h r o n o u s o r a s y n c h r o n o u s a c t i o n s a n d c a n d i f f r a l o n g m a n y o t h r d i m n s i o n s [ 1 1 ]. I n g n r a l, i n t r a c t i v c o m p u t a t i o n i n v o l v s, u n l i k a l g o r i t h m i c c o m p u t a t i o n, d y - n a m i c i n t r a c t i o n s t r a m s o f c o m p u t a t i o n a l a g n t s [ 1 1, 1 3 ]. S u c h a c o m p u t a t i o n c a n b c o n s i d r d a s a k i n d o f d i s t r i b u t d c o n t r o l o f i n f o r m a t i o n s t r a m s a n d a g n t s a c t i v i t i s i n a d y n a m i c n v i r o n m n t [ 1 2 ]. A c l u s t r o f a g n t s o f t n s h o w s a b h a v i o r t h a t i s r a t h r c o m p l x a n d o r g a n i z d, d s p i t t h s i m p l i c i t y o f a c h s i n g l a g n t. I n t h i s a s p c t, x p l i c i t a p p l i c a t i o n o f i n t r a c t i v c o m p u t a t i o n s l a d s f r o m s y s t m s w i t h a l g o - r i t h m i c b h a v i o r t o s y s t m s w i t h i t h r s q u n t i a l i n t r a c t i v b h a v i o r ( a p a i r o f i n - t r a c t i n g a g n t s ) o r w i t h m u l t i - i n t r a c t i v b h a v i o r. T h i s o b s r v a t i o n c o n f o r m s p r - f c t l y t o t h s o f t w a r n g i n r i n g p r i n c i p l s o f p r o g r a m m i n g - i n - t h - l a r g a n d p r o - g r a m m i n g - i n - t h - s m a l l. M o r o v r, t h a g n t - b a s d a p p r o a c h i n ( d o m a i n - s p c i f i c ) l a n g u a g d s i g n a n d i m p l m n t a t i o n s m s x t r m l y i n t r i g u i n g i n t h c o n t x t o f t h s o b s r v a t i o n s. 3 Attributd Automata A t t r i b u t d a u t o m a t a w r i n t r o d u c d a s a m o d l o f x c u t a b l s p c i f i c a t i o n s b a s d o n r g u l a r s t r u c t u r s w i t h a t t r i b u t s a t t a c h d t o s t r u c t u r n o d s. R g u l a r i t y i s h r t r a t d i n t r m s o f f o r m a l l a n g u a g s p r i m i t i v i t m s c a n b c o m p o s d i n t o a s t r u c t u r b y m a n s o f c o n c a t n a t i o n, s l c t i o n a n d i t r a t i o n o p r a t i o n s. A t t r i b u t s s r v f o r t h p r s n t a t i o n o f c o n t x t u a l r l a t i o n s, a s w l l a s o f p r o p r t i s a n d m a n i n g o f u n d r l y - i n g c o n c p t s. A t t r i b u t d a u t o m a t o n i s a s t a t t r a n s i t i o n m a c h i n w i t h d i s t r i b u t d f i n i t m m o r y a t i t s s t a t s a n d s p c i f i d a t i t s t r a n s i t i o n s c o m p u t a t i o n s a n d c o m m u n i c a t i o n a c t i o n s. I n t h i s a s p c t, a n a t t r i b u t d a u t o m a t o n i s s i m p l y a g n r a l i z a t i o n o f a t r a d i t i o n a l f i n i t a u t o m a t o n w i t h a t t r i b u t s a n d c o m p u t a t i o n a l r l a t i o n s a t t a c h d t o s t a t s a n d t r a n s i - t i o n s o f t h a u t o m a t o n, r s p c t i v l y. A t t r i b u t d a u t o m a t o n i n t r m s o f f o r m a l l a n -
Domain-S p c if ic L ang u ag A g nt s 8 5 g u ag s is c ons id r d as a r c og niz r b as d on r g u l ar d at a s t r u c t u r s, t h r s p c t iv c l as s of f or mal g r ammar s is t h at of r g u l ar at t r ib u t g r ammar s [ 1 7 ]. C ons id r x amp l s of s y nt ax -d ir c t d r c og niz r s f or b inar y nu mb r s ( F ig. 1 ) and f or a c ont x t -s ns it iv l ang u ag ( F ig. 2 ). A t t r ib u t d au t omat on ( F ig. 1 ) r c og niz s b inar y nu mb r s and t h f inal at t r ib u t v al u a r p r s nt s t h d c imal v al u of t h b inar y nu mb r. I n anot h r au t omat on ( F ig. 2 ) at t r ib u t s ar in a d if f r nt r ol, t h y c ol l c t c ont x t u al inf or mat ion u s d in s om s t at s t o s l c t t h n x t t r ans it ion. A s ou r x amp l s d mons t r at, t h r ar al t r nat iv p os s ib il it i s t o c ons t r u c t t r ad it ional s y n- t ax -d ir c t d c omp il r s on t h b as is of s imp l r g u l ar at t r ib u t d s t r u c t u r s. a := 0 a := 2a 0 a 1. a := a; b := 0; c := 0 a := a b := 2b + 1 c := c + 1 1 a, b, c 0 a := a b := 2b; c := c + 1 a := 2a + 1 a := a a a := a + b 2 -c Fig. 1. Rcognizr of binary numbrs. k := 0 k := k + 1 a b k := k c b k := k c l := l l := 1 m := 1 k k, l k, l, m k k := k l := l + 1 k := k l : = l m := m + 1 k = l = m k := k k := k Fig. 2. Rcognizr of t h cont x t -snsit iv l anguag L = { a n b n c n n 0 }. T h r x is t s v r al x t ns ions of t h c onc p t of f init s t at mac h in w it h m mor y. A t t r ib u t d au t omat on c an b d is t ing u is h d among t h m b y d is t r ib u t d m mor y, i.. b y al l oc at ing m mor y t o t h s t at s of t h c omp u t at ion. Dis t r ib u t d m mor y t og t h r w it h t h l oc al d f init ion of d at a t r ans f or mat ion f u nc t ions h l p s t o d c omp os / c omp os an at t r ib u t d au t omat on. N ot t h at h i r ar c h ic al c omp os it ion f u nc t ions as a t ool f or t h ad q u at mod l ing of h i r ar c h ic al d at a s t r u c t u r s and h i r ar c h ic al c omp u t at ional s t r u c t u r s. T h is id a is r oot d in t h int r p r t ing au t omat a c onc p t, ap p l i d in V i nna m t h od f or d f ining p r og r amming l ang u ag s [ 1 8 ]. I n t h mod l ing of int r ac t iv s y s t ms it ap p ar s imp or t ant f or t h s y s t m t o r ac t ad q u at l y t o t h c h ang s in it s nv ir onm nt. T h s c h ang s c annot b p r d ic t d, s u c h a s y s t m is inh r nt l y i n t r ac t i v, i.. r s p ond s t o c h ang s in it s nv ir onm nt b y p r f or ming i n t r n al c h ang s, w h ic h, in t u r n, w il l b r g is t r d b y t h nv ir onm nt as s om int r nal v nt s of t h s y s t m. W c al l A A s imu l at ing int r ac t iv s y s t ms i n t r - ac t i v A A. I n t h s au t omat a, t h s q u nc of int r nal v nt s ( t r ans f or mat ions of at -
86 M r i k M r i s t, T õ n i s K l d r, J ü r i H l k i v i, L o M o t u s t r i b u t v a l u s ) w i l l b i n s o m m a n n r s y n c h r o n i z d w i t h x t r n a l v n t s i n t h n v i - r o n m n t. I n t r a c t i v a t t r i b u t d a u t o m a t a r p r s n t a c r t a i n k i n d o f multi- s tr a m in - t r a ctio n ma ch in s i n t r o d u c d b y W g n r [ 1 1, 1 2, 1 3 ]. F o r x a m p l, t h p a r s i n g o f D y c k l a n g u a g s [ 1 9 ] i s s o l v d b y i n t r a c t i v A A a s f o l l o w s ( F i g. 3 ). T h r g u l a r s t r u c t u r o f a s t r i n g i s r p r s n t d b y t h m o v s o f A A, c o u n t r s o f p a r n t h s s a r r p r s n t d a s a t t r i b u t s. A n i n t r a c t i v a u t o m a t o n i s c o n s t r u c t d f o r c o u n t i n g c r t a i n k i n d s o f p a r n t h s s ( [ ]. ( ), { } ). A u t o m a t a i n t r - a c t w i t h a c h o t h r f o r r c o g n i t i o n o f a s t r i n g o f p a r n t h s s. n=0 ( n=n+1; c=c ( n, c ) n, c n=n-1; c=c&(n>0) ) [ [ n=n-1; c=c&(n>0) n=n+1; c=c c = A (c) [ ] c = A (c) [ ] c = c&(n=0) c a) Accounting of p ar nth s is '( ' and ')': c' = A ( ) ( c) n=0 [ n=n+1; c=c [ n, c ] n, c n=n-1; c=c&(n>0) ] ( ( n=n-1; c=c&(n>0) n=n+1; c=c ( ) c = A (c) ( ) c = A (c) c = c&(n=0) c b ) Accounting of p ar nth s is '[ ' and '] ': c' = A [ ] ( c) Fig. 3. G r am m ar : S > S S ( ) ( S ) [ ] [ S ] 4 Intractiv Modling of Languags T h c o n s i d r a t i o n s d s c r i b d a b o v l d u s t o t h d s i g n a n d i m p l m n t a t i o n o f a s y s t m f o r t h v i s u a l d v l o p m n t o f i n t r a c t i v A A. S u c h a d v l o p m n t n v i r o n - m n t w i l l s r v a s a n i n s t r u m n t a l t o o l f o r t h d s i g n, a s w l l a s f o r t h i m p l m n t a - t i o n o f v a r i o u s a p p l i d s o f t w a r. T o q u o t P t r W g n r [ 1 1 ] : T h n g a t i v r s u l t t h a t i n t r a c t i v b h a v i o r i s n o t x p r s s i b l b y T u r i n g m a c h i n s d t r m i n s a p o s i t i v c h a l l n g t o d v l o p p r a c t i c a l m o d l s o f i n t r a c t i v c o m p u t a t i o n. M o r o v r, w a r n c o u r a g d b y o u r p i l o t p r a c t i c x p r i n c. I n t r a c t i v A A c o m p o s i t i o n s a r x p r s - s i v i n s o l v i n g n o n - a l g o r i t h m i c p r o b l m s. S o m a l g o r i t h m i c p r o b l m s c a n b m o r f f i c i n t l y s o l v d b y i n t r a c t i v t c h n i q u s.
Domain-S p c if ic L ang u ag A g nt s 8 7 T h b as ic int r ac t iv A A c omp os it ion t c h niq u s ar imp l m nt d in a p r ot ot y p s y s t m p r og r amm d in J av a. T h c ommu nic at ion t c h niq u ap p l i d is t h at of J av a M s s ag S r v ic ( J M S ). A u t omat a ar int r ac t ing b y s nd ing m s s ag s, b ot h p oint - t o-p oint and p u b l is h -and -s u b s c r ib m t h od s ar av ail ab l. S nd ing a m s s ag is t r at d as a s p ar at p r oc s s t h at c an af f c t t h v al u at ion of at t r ib u t s f or t h n x t s t at. A m s s ag c an b ac c p t d it h r in t h s y nc h r onou s mod ac c p t anc inc l u d d in t h s l c t ion p r oc s s of t h n x t s t at or, in t h as y nc h r onou s mod an ar r iv ing m s s ag init iat s a s p c if ic s p ar at p r oc s s of ac c p t anc, w h ic h in it s ow n t u r n may af f c t t h v al u s of at t r ib u t s. T h c ol l c t ion of int r ac t iv au t omat a d s ig n d t o s ol v a p r ob l m f or ms a c l u s t r of ag nt s, i.. a mu l t i-ag nt. M mb r ag nt s of a c l u s t r may b ac t iv at d as on c omp l t t as k on a c omp u t r or, as a d is t r ib u t d t as k. I n t h f ir s t c as, ag nt s c an x - c h ang inf or mat ion b y c ommon ( c l u s t r ) at t r ib u t s and m s s ag s, in t h s c ond c as onl y b y m s s ag s. A c l u s t r of ag nt s of t n s h ow s a b h av ior t h at is r at h r c omp l x and or g aniz d, d s p it t h s imp l ic it y of ac h s ing l ag nt. I n t h is as p c t, ap p l ic at ion of int r ac t iv A A ( as int r ac t ion mac h in s ) l ad s f r om s y s t ms w it h al g or it h mic b - h av ior t o s y s t ms w it h it h r s q u nt ial int r ac t iv b h av ior ( a p air of int r ac t ing ag nt s ) or w it h mu l t i-int r ac t iv b h av ior. T h is ob s r v at ion c onf or ms p r f c t l y t o t h s of t w ar ng in r ing p r inc ip l s of p r og r amming -in-t h -l ar g and p r og r amming -int h -s mal l. M or ov r, t h ag nt -b as d ap p r oac h in ( d omain-s p c if ic ) l ang u ag s d s ig n and imp l m nt at ion s ms x t r m l y int r ig u ing in t h c ont x t of t h s ob s r v at ions. O u r p r ot ot y p s y s t m of int r ac t iv A A ( an ag nt -b as d p r ob l m-s ol v ing nv ir onm nt ) is, in a s ns, a s y s t m of p r og r amming w h r t h s of t w ar is c ons t r u c t d b y m ans of c ol l ab or at iv ( int r ac t ing ) c omp u t at ional ag nt s. C omp on nt s c ons t r u c t d ar s av d as it ms of t h c ommon d at ab as of au t omat a ag nt s. F r om t h s p c if ic a- t ion of ag nt s of a p ar t ic u l ar t as k, a J av a-p r og r am w il l b c omp il d. N ot ions and t r ms ap p l i d in an au t omat on ar s p c if i d b y m ans of t h s o-c al l d axiom-c l as s s ( s p c if ic J av a-c l as s s, r p r s nt ing not ions of t h p r ob l m d omain). N ot ions ar imp l m nt d as p ar t ic u l ar n ot ion -c l as s s d r iv d f r om ax iom-c l as s s. T h ap p l ic at ion is d r iv d f r om p ar t ic u l ar not ion-c l as s s and c omp il d in t h c ont x t of c ons t r u c t s ( t r ms and not ions ), s p c if i d b y t h u s r. S u c h a s t y l of imp l m nt at ion s u p p or t s t h s y s t m s f l x ib il it y b y c h ang s in not ion-c l as s s n w p r op r t i s c an b int r od u c d t o t h c l u s t r as a w h ol. O n t h ot h r h and, at s om l v l, J av a-p r og r amming is n d d. F r om t h v i w p oint of l ang u ag imp l m nt at ion w t ak a not ion v i w of DS L d s ig n, in t h at a l ang u ag is d s ig n d as a s t of int r r l at d au t onomou s not ions. A l ang u ag p r oc s s or w il l b c ons t r u c t d as a c l u s t r of int r ac t iv A A ( ag nt s ) of l ang u ag not ions. A n ag nt " r p r s nt s " an ins t anc of a p ar t ic u l ar not ion, i.. t h not ion' s r p r s nt at ion, it s s t r u c t u r, p r op r t i s and m aning. T h t as k of t h not ion ag nt is t o s c u r t h ap p r op r iat t r ans l at ion/ int r p r t at ion of v r y not ion ins t anc in it s g iv n s p c if ic c ont x t. T h c ont x t u al and s t r u c t u r al r l at ions of t h not ions inc l u d d in a l ang u ag ar s p c if i d in t r ms of t h p r op r t i s of t h not ion ag nt and it s int r ac t ions w it h ot h r not ion ag nt s. I f n c s s ar y, t h not ion ag nt s w il l ap p l y ot h r ag nt s f or t r ad it ional s u b t as k s of s y nt ax -d ir c t d t r ans l at ion and c od g n r at ion. I n ot h r w or d s, an imp l m nt at ion of a l ang u ag is s p c if i d as a mu l t i-s t r am int r ac - t ion mac h in [ 1 2, 1 3 ]
88 M r i k M r i s t, T õ n i s K l d r, J ü r i H l k i v i, L o M o t u s A s a n x a m p l o f t h m u l t i - a g n t a p p r o a c h f o r D S L i n t h f r a m w o r k o f i n t r a c - t i v a t t r i b u t d a u t o m a t a, l t u s c o n s i d r a t i n y i n t r a c t i o n l a n g u a g f o r a n o n l i n t i c k t s a l s s y s t m. T h x a m p l p r o b l m i s b o r r o w d f r o m [ 2 0 ]. C u s t o m r s b u y t i c k t s f r o m a t i c k t s r v r. T h s r v r c o m m u n i c a t s o n l y w i t h s a l s a g n t s. C u s t o m r s c a n a s k t h a g n t f o r v a r i o u s s r v i c s. T h s s r v i c s i n c l u d r s r v i n g t i c k t s, p a y i n g f o r a n d g t t i n g t h r s r v d t i c k t s, a n d c a n c l i n g r s r v a t i o n s. T h l a n g u a g o f c o m m u - n i c a t i o n b t w n t h c u s t o m r s a n d a g n t s c o n s i s t s o f a c o u p l o f n o t i o n s o n l y, a s g i v n i n f i g u r 4. I m p l m n t a t i o n o f t h l a n g u a g b y a u t o m a t a i s g i v n i n f i g u r 5. Agnt Customr Tickt srvr Agnt Customr Customr > Agnt: want to buy, cancl_rs, pay Agnt > Customr: rsrvd, not approvd, cancld, tickts, sold_out, try_latr, too_many, quu_full Fig. 4. Agnt tickts sold out mssag not undrstood mssag not undrstood Wait inmssag pc. 1 rsrv Rqust inmssag pc. 2 tickts rsrvd Pay inmssag accpt tickts or rjct Ship rjct pc. 3 or pay inmssag pc. 4 mssag not undrstood mssag not undrstood Prcondition 1: Wait srvr failur ==> Wait want to buy ==> Rqust othrwis ==> Wait Prcondition 2: Rqust tickts availabl ==> Pay tickts sold out ==> Wait othrwis ==> Rqust Prcondition 3: Pay rqust accptd ==> Ship rqust rjctd ==> Ship othrwis ==> Pay Prcondition 4: Ship don ==> Wait othrwis ==> Ship Fig. 5.
Domain-S p c if ic L ang u ag A g nt s 8 9 5 Conclusion t l t c t p f s s t s s t g l s t f t g t t t p s t s p T c u c t p c t l s t s s t t t t f t s u p p t t l s t u c t u t p l t s p f l f k f t s l ( p c l u s t t l l s l t c s l l l s f p c l u s A s t l u s, t L, t f s t p t l s v v g l t p s s c s v A s t l, l t t t s t u c t u s s t t t p s s s s c s c l t t l t s s t l p f t f l t s t t l Du r ing h as d ad h ar adig m or d ig ning of w ar y ms h as r adu al y h if d r om h al or it h m-c nt r d o h int r ac ion-c nt r d ap r oac h. I d a of ag nt -or i nt d of w ar ng in r ing is r ap idl y r ading. h r ial m h odol og ic al as of mod ing of w ar y ms is o w h at x nt h or mal m h ods or h r as onab r r ing of h r ob m domain. I nt r ac iv ag nt ap ar as a as ib b as ic r am w or or h d ig n and imp m nt at ion of domain-s if ic ) ang ag. I nt r ac iv r ans at ion mod nab o r at a mant ic al y r ic h r mod d w or d, al o or domain-s if ic ang ag. any ot h r ar if ic ial ang ag h DS oo, ar ir onl y ar ial y d ig n d, ol ing r adu al y in h r oc of ommu nic at ion and ob r at ion. an int r ac iv mod it ads o b r r r d y m, b r x r iv n and r du omp x it y. O n h ot h r h and, h inc omp n of h mod is inh r nt r ic or h r dom of mu i-ag nt d ig n, i.. it r mains an ar of r ans at ion R f r nc s A A : R S L : z D ( a L S 0 S a R R S A S A ) A A a ( A A A R z a 0 G A A R D S a A S R G S A A ( ) 0 G a D D S A a D D S R R R A 0 ( ) 0 S : a a R G D : ( ) a a L S 1. M r i s t, M., P n j a m, J.: t t r i b u t d F i n i t u t o m a t a. I n Proc. of Int. Workshop CC 9 2 on Com pil r-com pil r, p o r t s o f t h U n i v r s i t y o f P a d r b o r n 10 3, 19 9 3, p p.4 8 5 1. 2. M r i s t, M., P n j a m, J.: T o w a r d K n o w l d g - b a s d p c i f i c a t i o n s o f a n g u a g s. I n J.B a r d i n s,.b j o r n r E d s.), B l tic Com pu t r S ci nc, N C, 5 2, p r i n g r V r l a g, 19 9 1, p p.6 5 7 6 3. M r i s t, M., P n j a m, J.: A ttrib u t d F init A u tom ta. s. p. C 2 3 / 9 1, I n s t i t u t o f C y b r n t i c s, E s t o n i a n c a d m y o f c i n c s, T a l l i n n, 19 9 1, 15 p. 4. M r i s t, M., P n j a m, J.: t t r i b u t d M o d l s o f C o m p u t i n g. Proc d ing s of th E stonia n A ca d m y of S ci nc s. E ng in ring, 1( 2, 19 9 5, p p. 13 9 15 7. 5. M r i s t, M., P n j a m, J., V n, V.: M o d l s o f t t r i b u t d u t o m a t a. Inform tica, 9 1), 19 9 8, p p.8 5 10 5. 6. J u h o l a, M., M r i s t, M.: n t t r i b u t d u t o m a t o n f o r c o g n i i n g o f N y s t a g m u s E y M o v m n t s. IA PR Pa p rs on S tru ctu ra l nd S y nta ctic Pa tt rn R cog nition, B r n, 19 9 3, p p.19 4 2 6. 7. r ö n f o r s, T., M r i s t, M.: t t r i b u t d u t o m a t a i n P a t t r n c o g n i t i o n o f i g i t a l i g n a l s. Com pu t r M thod s nd Prog ra m s in B iom d icin, 19 9 3, p p.7 6 3 7 8 5. 8. K o s k i,., J u h o l a, M., M r i s t, M.: y n t a c t i c c o g n i t i o n o f E C i g n a l s b y t t r i b u t d F i n i t u t o m a t a. Pa tt rn R cog nition, 2 8 12, 19 9 5, p p.19 2 7 19 4. 9. r ö n f o r s, T.: N ov l M thod s of S y nta ctic Pa tt rn R cog nition for P k D t ction of A u d itory B ra inst m R spons s. o c t o r a l i s s r t a t i o n, U n i v r s i t y o f K u o p i o, P u b l i c a t i o n s C. N a t u r a l a n d E n v i r o n m n t a l c i n c s 2 8, 19 9 4. 10. K o s k i,.: O n S tru ctu ra l R cog nition nd A na l y sis M thod s A ppl i d to E CG S ig na l s. o c t o r a l i s s r t a t i o n, U n i v r s i t y o f T u r k u, C o m p u t r c i. s. p o r t s - 9 7-1, 19 9 7. 11. W g n r, P.: W h y I n t r a c t i o n i s m o r P o w r f u l t h a n l g o r i t h m s. Com m u nica tions of th A CM, 4 5, 19 9 7, p p.8 9 1. 12. W g n r, P.: I n t r a c t i v o f t w a r T c h n o l o g y. I n H nd b ook of Com pu t r S ci nc nd E ng in ring, C C P r s s, 19 9 7. 13. W g n r, P., o l d i n,.: I n t r a c t i o n a s a F r a m w o r k f o r M o d l l i n g. I n C h n t a l d s. Conc ptu l M od l l ing : Cu rr nt Issu s nd F u tu r D ir ctions, 19 9 9, N C v o l. 15 6 5.
90 M r i k M r i s t, T õ n i s K l d r, J ü r i H l k i v i, L o M o t u s 14. W o o l d r i d g, M., J n n i n g s, N.R.: I n t l l i g n t a g n t s : t h o r y a n d p r a c t i c. Knowl d g E ng i - n r i ng R v i w, 10 ( 2 ), 19 9 5, p p.115 15 2. 15. F r b r, J.: M u lt i -A g nt S y s t m s. A d d i s o n - W s l y, 19 9 9. 16. W n g r, E.: C om m u ni t i s of P r a c t i c. C a m b r i d g U n i v r s i t y P r s s, L o n d o n, 19 9 8. th 17. M r i s t, M., V n, V.: A t t r i b u t d A u t o m a t a a n d L a n g u a g R c o g n i z r s. I n : P r oc. of 4 S y m p os i u m on P r og r a m m i ng L a ng u a g s a nd S of t wa r T ools. V i s g r á d, H u n g a r y, J u n 9 10, 19 9 5, p.114 12 1. 18. O l l o n g r n, A.. D f i n i t i o n o f P r o g r a m m i n g L a n g u a g s b y I n t r p r t i n g A u t o m a t a. A c a d m i c P r s s, L o n d o n, 19 7 4. 19. G i n s b u r g, C., G r i b a c h, S.: D t r m i n i s t i c c o n t x t - f r l a n g u a g s. I nf or m. a nd C ont r ol, 9 ( 6 ), 19 9 6, p p.6 2 0 6 48. 2 0. W a n g, W., H i d v é g i, Z., B a i l y J r., A.D., W h i n s t o n, A.B.: E - P r o c s s D s i g n a n d A s s u r a n c U s i n g M o d l C h c k i n g. C om p u t r, 3 3 ( 10 ), 2 0 0 0, p p.48 5 3.