A Strong Convergence Theorem for a Proximal-Type. Algorithm in Re exive Banach Spaces

Size: px
Start display at page:

Download "A Strong Convergence Theorem for a Proximal-Type. Algorithm in Re exive Banach Spaces"

Transcription

1 A Strog Covergece Theorem for a Proxmal-Type Algorthm Re exve Baach Spaces Smeo Rech ad Shoham Sabach Abstract. We establsh a strog covergece theorem for a proxmal-type algorthm whch approxmates (commo) zeroes of maxmal mootoe operators re exve Baach spaces. Ths algorthm employs a well-chose covex fucto. The behavor of the algorthm the presece of computatoal errors ad the case of zero free operators s also aalyzed. Fally, we meto several corollares, varatos ad applcatos. 1. Itroducto I ths paper X deotes a real re exve Baach space wth orm kk ad X stads for the (topologcal) dual of X edowed wth the duced orm kk. We deote the value of the fuctoal 2 X at x 2 X by h; x. A operator A : X! 2 X s sad to be mootoe f for ay x; y 2 dom A, we have 2 Ax ad 2 Ay =) h ; x y 0: (Recall that the set dom A = fx 2 X : Ax 6=?g s called the e ectve doma of such a operator A.) A mootoe operator A s sad to be maxmal f graph A, the graph of A, s ot a proper subset of the graph of ay other mootoe operator. I ths paper f : X! ( 1; +1] s always a proper, lower semcotuous ad 2000 Mathematcs Subject Class cato. Prmary: 47H05; Secodary: 47J25. Key words ad phrases. Baach space, Bregma projecto, Legedre fucto, maxmal mootoe operator, mootoe operator, proxmal pot algorthm, resolvet, totally covex fucto. 1

2 2 SIMEON REICH AND SHOHAM SABACH covex fucto, ad f : X! ( 1; +1] s the Fechel cojugate of f. The set of oegatve tegers wll be deoted by N. The problem of dg a elemet x 2 X such that 0 2 Ax s very mportat Optmzato Theory ad related elds. For example, f A s the subd of f, the A s a maxmal mootoe operator ad the equato 0 (x) s equvalet to the problem of mmzg f over X. Oe of the methods for solvg ths problem Hlbert space s the well-kow proxmal pot algorthm. Let H be a Hlbert space ad let I deote the detty operator o H. The proxmal pot algorthm geerates, for ay startg pot x 0 = x 2 H, a sequece fx g 2N H by the rule (1.1) x +1 = (I + A) 1 x ; = 0; 1; 2; : : : ; where f g 2N s a gve sequece of postve real umbers. Note that (1.1) s equvalet to 0 2 Ax (x +1 x ) ; = 0; 1; 2; : : : : Ths algorthm was rst troduced by Martet [26] ad further developed by Rockafellar [34], who proves that the sequece geerated by (1.1) coverges weakly to a elemet of A 1 (0) whe A 1 (0) s oempty ad lm f!+1 > 0. Furthermore, Rockafellar [34] asks f the sequece geerated by (1.1) coverges strogly. Ths questo was aswered the egatve by Güler [22], who preseted a example of a subd eretal for whch the sequece geerated by (1.1) coverges weakly but ot strogly; see [6] for a more recet ad smpler example. More recetly, Solodov ad Svater [37] have mod ed the proxmal pot algorthm order to geerate a strogly coverget sequece. They troduce the followg algorthm:

3 PROXIMAL-TYPE ALGORITHM 3 (1.2) 8 >< >: x 0 2 H; 0 = v + 1 (y x ) ; v 2 Ay ; H = fz 2 H : hv ; z y 0g ; W = fz 2 H : hx 0 x ; z x 0g ; x +1 = P H\W (x 0 ); = 0; 1; 2; : : : : Here, for each x 2 H ad each oempty, closed ad covex subset C of H, the mappg P C s de ed by kx P C xk = f fkx zk : z 2 Cg. Ths mappg s called the metrc projecto of H oto C. They prove that f A 1 (0) s oempty ad lm f!+1 > 0, the the sequece geerated by (1.2) coverges strogly to P A 1 (0). Kammura ad Takahash [25] geeralze ths result to those Baach spaces X whch are both uformly covex ad uformly smooth. They troduce the followg algorthm: 8 x 0 2 X; (1.3) >< >: 0 = v + 1 (Jy Jx ) ; v 2 Ay ; H = fz 2 X : hv ; z y 0g ; W = fz 2 X : hjx 0 Jx ; z x 0g ; x +1 = Q H\W (x 0 ); = 0; 1; 2; : : : ; where J s the ormalzed dualty mappg of the space X. Here, for each oempty, closed ad covex subset C of X, Q C s a certa geeralzato of the metrc projecto P C H. They prove that f A 1 (0 ) s oempty ad lm f!+1 > 0, the the sequece geerated by (1.3) coverges strogly to Q A 1 (0 ). Other developmets regardg the proxmal pot algorthm ca be foud, for example, [4, 6, 9, 10, 14, 16, 18, 20, 23, 24, 27, 29, 30, 31, 35, 38]. I the preset paper we study a exteso of algorthms (1.2) ad (1.3) to all re exve Baach spaces usg a well-chose covex fucto f. More precsely, we

4 4 SIMEON REICH AND SHOHAM SABACH cosder the followg algorthm troduced by Gárcga Otero ad Svater [21]: 8 x 0 2 X; (1.4) >< >: 0 = + 1 (rf(y ) rf(x )) ; 2 Ay ; H = fz 2 X : h ; z y 0g ; W = fz 2 X : hrf(x 0 ) rf(x ); z x 0g ; x +1 = proj f H \W (x 0 ); = 0; 1; 2; : : : ; where f g 2N s a gve sequece of postve real umbers, rf s the gradet of f ad proj f C s the Bregma projecto (see Secto 2.4) of X oto C duced by f. Algorthm (1.4) s more exble tha (1.3) because t leaves us the freedom of ttg the fucto f to the ature of the operator A (especally whe A s the subd eretal of some fucto) ad of the space X ways whch make the applcato of (1.4) smpler tha that of (1.3). It should be observed that f X s a Hlbert space H, the usg (1.4) the fucto f (x) = (1=2) kxk 2, oe obtas exactly algorthm (1.2). If X s ot a Hlbert space, but stll a uformly covex ad uformly smooth Baach space X, the settg f (x) = (1=2) kxk 2 (1.4), oe obtas exactly (1.3). We also ote that the choce f (x) = (1=2) kxk 2 some Baach spaces may make the computatos algorthm (1.3) qute d cult. These computatos ca be smpl ed by a approprate choce of f. For stace, f X = `p or X = L p wth p 2 (1; +1), ad f (x) = (1=p) kxk p (1.4), the the computatos become smpler tha those requred (1.3), whch correspods to f (x) = (1=2) kxk 2. We propose a exteso of algorthm (1.4) (see algorthm (3.1)) whch approxmates a commo zero of several maxmal mootoe operators ad whch allows computatoal errors. Our ma result (Theorem 1) s formulated ad proved Secto 3. The ext secto s devoted to several prelmary de tos ad results. The behavor of the algorthm whe the operator A s zero free s aalyzed Secto 4 (see Theorem 2). The fth secto cotas three corollares of Theorems 1 ad 2. I the sxth ad last secto we preset a applcato of Theorems 1 ad 2.

5 PROXIMAL-TYPE ALGORITHM 5 2. Prelmares 2.1. Some facts about Legedre fuctos. Legedre fuctos mappg a geeral Baach space X to ( 1; +1] are de ed [3]. Accordg to [3, Theorems 5.4 ad 5.6], sce X re exve, the fucto f s Legedre f ad oly f t sats es the followg two codtos: (L1) The teror of the doma of f, t dom f, s oempty, f s Gâteaux d eretable (see below) o t dom f ad dom rf = t dom f; (L2) The teror of the doma of f, t dom f, s oempty, f s Gâteaux d eretable o t dom f ad dom rf = t dom f : Sce X s re exve, we always have (@f) 1 (see [7, p. 83]). Ths fact, whe combed wth codtos (L1) ad (L2), mples the followg equaltes: rf = (rf ) 1 ; ra rf = dom rf = t dom f ad ra rf = dom rf = t dom f: Also, codtos (L1) ad (L2), cojucto wth [3, Theorem 5.4], mply that the fuctos f ad f are strctly covex o the teror of ther respectve domas. Several terestg examples of Legedre fuctos are preseted [2] ad [3]. Amog them are the fuctos 1 s kks wth s 2 (1; 1), where the Baach space X s smooth ad strctly covex ad, partcular, a Hlbert space. The fucto f s called co te f dom f = X.

6 6 SIMEON REICH AND SHOHAM SABACH 2.2. A property of gradets. For ay covex f : X! ( 1; +1] we deote by dom f the set fx 2 X : f (x) < +1g. For ay x 2 dom f ad y 2 X, we deote by f (x; y) the rght-had dervatve of f at x the drecto y, that s, f f(x + ty) (x; y) := lm t&0 t f(x) : The fucto f s sad to be Gâteaux d eretable at x f lm t!0 (f(x + ty) f(x)) =t exsts for ay y. The fucto f s sad to be Fréchet d eretable at x f ths lmt s attaed uformly kyk = 1. Fally, f s sad to be uformly Fréchet d eretable o a subset E of X f the lmt s attaed uformly for x 2 E ad kyk = 1. We wll eed the followg result. Proposto 1. If f : X! R s uformly Fréchet d eretable ad bouded o bouded subsets of X, the rf s uformly cotuous o bouded subsets of X from the strog topology of X to the strog topology of X. Proof. If ths result were ot true, there would be bouded sequeces fx g 2N ad fy g 2N, ad a postve umber " such that kx y k! 0 ad hrf (x ) rf (y ) ; w 2", where fw g 2N s a sequece X wth kw k = 1 for each 2 N. Sce f s uformly Fréchet d eretable, there s a postve umber such that f (y + tw ) f (y ) t hrf (y ) ; w "t for all 0 < t < ad 2 N. We also have hrf (x ) ; (y + tw ) x f (y + tw ) f (x ) ; 2 N: I other words, t hrf (x ) ; w f (y + tw ) f (y ) + hrf (x ) ; x y + f (y ) f (x ) :

7 PROXIMAL-TYPE ALGORITHM 7 Hece 2"t t hrf (x ) rf (y ) ; w [f (y + tw ) f (y ) t hrf (y ) ; w ] + hrf (x ) ; x y + f (y ) f (x ) "t + hrf (x ) ; x y + f (y ) f (x ) : Sce rf s bouded o bouded subsets of X (see [12, Proposto , p. 17]), t follows that hrf (x ) ; x y coverges to zero, whle f (y ) f (x )! 0 sce f s uformly cotuous o bouded subsets (see [1, Theorem 1.8, p. 13]). But ths would yeld that 2"t "t, a cotradcto Some facts about totally covex fuctos. Let f : X! ( 1; +1] be covex. The fucto D f : dom f t dom f! [0; +1] de ed by D f (y; x) := f(y) f(x) f (x; y x); s called the Bregma dstace wth respect to f (cf. [17]). If f s a Gâteaux d eretable fucto, the the Bregma dstace has the followg mportat property, called the three pot detty: for ay x; y; z 2 t dom f, (2.1) D f (x; y) + D f (y; z) D f (x; z) = hrf(z) rf(y); x y : Recall that, accordg to [12, Secto 1.2, p. 17] (see also [11]), the fucto f s called totally covex at a pot x 2 t dom f f ts modulus of total covexty at x, that s, the fucto f : t dom f [0; +1)! [0; +1], de ed by (2.2) f (x; t) := f fd f (y; x) : y 2 dom f; ky xk = tg ; s postve wheever t > 0. The fucto f s called totally covex whe t s totally covex at every pot x 2 t dom f. I addto, the fucto f s called totally covex o bouded sets f f (E; t) s postve for ay oempty bouded subset E of X ad for ay t > 0, where the modulus of total covexty of the fucto f o

8 8 SIMEON REICH AND SHOHAM SABACH the set E s the fucto f : t dom f [0; +1)! [0; +1] de ed by f (E; t) := f f f (x; t) j x 2 E \ dom fg : Examples of totally covex fuctos ca be foud, for example, [12, 15]. The followg proposto summarzes some propertes of the modulus of total covexty. Proposto 2 (cf. [12, Proposto 1.2.2, p. 18]). Let f be a proper, covex ad lower semcotuous fucto. If x 2 t dom f, the () The doma of f (x; ) s a terval of the form [0; f (x)) or [0; f (x)] wth f (x) 2 (0; +1]. () If c 2 [1; +1) ad t 0, the f (x; ct) c f (x; t). () The fucto f (x; ) s superaddtve, that s, for ay s; t 2 [0; +1), we have f (x; s + t) f (x; s) + f (x; t). (v) The fucto f (x; ) s creasg; t s strctly creasg f ad oly f f s totally covex at x. The followg proposto follows from [14, Proposto 2.3, p. 39] ad [39, Theorem , p. 164]. Proposto 3. If f s Fréchet d eretable ad totally covex, the f s co te. The ext proposto turs out to be very useful the proof of our ma result. Proposto 4 (cf. [32, Proposto 2.2, p. 3]). If x 2 dom f, the the followg statemets are equvalet: () The fucto f s totally covex at x; () For ay sequece fy g 2N dom f, lm D f (y ; x) = 0 ) lm ky xk = 0:!+1!+1

9 PROXIMAL-TYPE ALGORITHM 9 Recall that the fucto f s called sequetally cosstet (see [15]) f for ay two sequeces fx g 2N ad fy g 2N X such that the rst oe s bouded, lm D f (y ; x ) = 0 ) lm ky x k = 0:!+1!+1 Proposto 5 (cf. [12, Lemma 2.1.2, p. 67]). If dom f cotas at least two pots, the the fucto f s totally covex o bouded sets f ad oly f the fucto f s sequetally cosstet The resolvet of A relatve to f. Let A : X! 2 X be a operator ad assume that f Gâteaux d eretable. The operator Prt f A := (rf + A) 1 : X! 2 X s called the protoresolvet of A, or, more precsely, the protoresolvet of A relatve to f. Ths allows us to de e the resolvet of A, or, more precsely, the resolvet of A relatve to f, troduced ad studed [5], as the operator Res f A : X! 2X gve by Res f A := Prtf A rf. Ths operator s sgle-valued whe A s mootoe ad f s strctly covex o t dom f. If A where ' s a proper, lower semcotuous ad covex fucto, the we deote Prox f ' := Prt ad proxf ' := Res : If C s a oempty, closed ad covex subset of X, the the dcator fucto C of C, that s, the fucto 8 >< 0 f x 2 C C (x) := >: +1 f x =2 C s proper, covex ad lower semcotuous, ad C exsts ad s a maxmal mootoe operator wth doma C. The operator prox f C s called the Bregma projecto oto C wth respect to f (cf. [8]) ad we deote t by proj f C. Note that f X s a Hlbert space ad f(x) = 1 2 kxk2, the the Bregma projecto of x oto C,.e., argm fky xk : y 2 Cg, s the metrc projecto P C.

10 10 SIMEON REICH AND SHOHAM SABACH Recall that the Bregma projecto of x oto the oempty, closed ad covex set K dom f, s the ecessarly uque vector proj f K (x) 2 K satsfyg D f proj f K (x); x = f fd f (y; x) : y 2 Kg : Smlarly to the metrc projecto Hlbert spaces, Bregma projectos wth respect to totally covex ad d eretable fuctos have varatoal characterzatos. Proposto 6 (cf. [15, Corollary 4.4, p. 23]). Suppose that f s totally covex o t dom f. Let x 2 t dom f ad let K t dom f be a oempty, closed ad covex set. If ^x 2 K, the the followg codtos are equvalet: () The vector ^x s the Bregma projecto of x oto K wth respect to f; () The vector ^x s the uque soluto of the varatoal equalty hrf (x) rf (z) ; z y 0; 8y 2 K; () The vector ^x s the uque soluto of the equalty D f (y; z) + D f (z; x) D f (y; x) ; 8y 2 K: 3. A Strog Covergece Theorem for a Proxmal-Type Algorthm I ths secto we study the followg algorthm whe Z := T N =1 A 1 (0 ) 6=?: 8 x 0 2 X; = + 1 rf(y ) rf(x ) ; 2 A y ; (3.1) >< H = z 2 X : ; z y 0 ; H := \ N =1 H ; W = fz 2 X : hrf(x 0 ) rf(x ); z x 0g ; >: x +1 = proj f H \W (x 0 ); = 0; 1; 2; : : : ;

11 where, for each = 1; 2; : : : ; N, ad 2N PROXIMAL-TYPE ALGORITHM 11 2N s a gve sequece of postve real umbers s the sequece of errors correspodg to the approxmate solutos of the resolvet equato. Note that f = 0, the y = Res f A (x ) : Theorem 1. Let A : X! 2 X, = 1; 2; : : : ; N, be N maxmal mootoe operators such that Z := T N =1 A 1 (0 ) 6=?. Let f : X! R be a Legedre fucto whch s bouded, uformly Fréchet d eretable ad totally covex o bouded subsets of X. The, for each x 0 2 X, there are sequeces fx g 2N whch satsfy (3.1). If, for each = 1; 2; : : : ; N, lm f!+1 > 0, ad the sequeces of errors 2N X satsfy lm!+1 = 0 ad lm sup!+1 ; y 0, the each such sequece fx g 2N coverges strogly to proj f Z (x 0) as! +1. Proof. Note that dom rf = X because dom f = X ad f s Legedre. Hece t follows from [5, Corollary 3.14(), p. 606] that dom Res f A = X. We beg wth the followg clam. Clam 1: There are sequeces fx g 2N whch satsfy (3.1). As a matter of fact, we wll prove that, for each x 0 2 X, there exsts a sequece fx g 2N whch s geerated by (3.1) wth = 0 for all = 1; 2; : : : ; N ad 2 N. It s obvous that H are closed ad covex sets for ay = 1; 2; : : : ; N. Hece H s also closed ad covex. It s also obvous that W s a closed ad covex set. Let u 2 Z. Sce dom Res f 0 A = X, there exsts y 0; 0 2 X X such that 0 = rf(y 0) rf(x 0 ) y0 = Res f 0 (x 0) ad 0 A 0 2 A y0. Sce A s mootoe, t follows that 0 ; y 0 u 0; whch mples that u 2 H 0. Sce ths holds for ay = 1; 2; : : : ; N, t follows that u 2 H 0. It s also obvous that u 2 W 0 = X. Thus u 2 H 0 \ W 0, ad therefore x 1 = proj f H 0\W 0 (x 0 ) s well de ed. Now suppose that u 2 H 1 \ W 1 for some

12 12 SIMEON REICH AND SHOHAM SABACH 1. Let x = proj f H 1\W 1 (x 0 ). Aga, there exsts y ; 2 X X such that 0 = + 1 rf(y ) rf(x ) ad 2 Ay. The mootocty of A mples that u 2 H. Sce ths holds for ay = 1; 2; : : : ; N, t follows that u 2 H. Now t follows from Proposto 6() that hrf(x 0 ) rf(x ); u x D E = rf(x 0 ) rf proj f H 1\W 1 (x 0 ) ; u proj f H 1\W 1 (x 0 ) 0; whch mples that u 2 W. Therefore u 2 H \W, ad hece x +1 = proj f H \W (x 0 ) s well de ed. Thus the sequece we costructed s deed well de ed ad sats es (3.1), as clamed. From ow o we x a arbtrary sequece fx g 2N satsfyg (3.1). It s clear from the proof of Clam 1 that Z H \ W for each 2 N. Clam 2: The sequece fx g 2N s bouded. It follows from the de to of W ad Proposto 6() that proj f W (x 0 ) = x. Furthermore, by Proposto 6(), for each u 2 Z, we have (3.2) D f (x ; x 0 ) = D f proj f W (x 0 ); x 0 D f (u; x 0 ) D f u; proj f W (x 0 ) D f (u; x 0 ) : Hece the sequece fd f (x ; x 0 )g 2N s bouded by D f (u; x 0 ) for ay u 2 Z. Therefore the sequece f f (x 0 ; kx x 0 k)g 2N s bouded by D f (u; x 0 ), because from the de to of the modulus of total covexty (see (2.2)) ad from (3.2) we get that (3.3) f (x 0 ; kx x 0 k) D f (x ; x 0 ) D f (u; x 0 ):

13 PROXIMAL-TYPE ALGORITHM 13 Sce the fucto f s totally covex, the fucto f (x; ) s strctly creasg ad postve o (0; 1) (cf. Proposto 2(v)). Ths mples, partcular, that f (x; 1) > 0 for all x 2 X. Now suppose by way of cotradcto that the sequece fx g 2N s ot bouded. The there exsts a sequece f k g k2n of postve real umbers such that lm kx k k = +1: k!+1 Cosequetly, lm k!+1 kx k x 0 k = +1. Ths shows that the sequece f f (x 0 ; kx x 0 k)g 2N s ot bouded. Ideed, there exsts some k 0 > 0 such that kx k x 0 k > 1 for ay k > k 0 ad the, by Proposto 2(), we see that f (x 0 ; kx k x 0 k) kx k x 0 k f (x 0 ; 1)! +1; because, as oted above, f (x 0 ; 1) > 0. Ths cotradcts (3.3). Hece the sequece fx g 2N s deed bouded, as clamed. Clam 3: Every weak subsequetal lmt of fx g 2N belogs to Z. It follows from the de to of W ad Proposto 6() that proj f W (x 0 ) = x. Sce x +1 2 W, t follows from Proposto 6() that D f x +1 ; proj f W (x 0 ) + D f proj fw (x 0 ); x 0 D f (x +1 ; x 0 ) ad hece (3.4) D f (x +1 ; x ) + D f (x ; x 0 ) D f (x +1 ; x 0 ) : Therefore the sequece fd f (x ; x 0 )g 2N s creasg ad sce t s also bouded (see Clam 2), lm!+1 D f (x ; x 0 ) exsts. Thus from (3.4) t follows that (3.5) lm!+1 D f (x +1 ; x ) = 0:

14 14 SIMEON REICH AND SHOHAM SABACH Proposto 5 ow mples that lm!+1 (x +1 x ) = 0. For ay = 1; 2; : : : ; N, t follows from the three pot detty (see (2.1)) that D f (x +1 ; x ) D f y ; x = D f x +1 ; y + rf(x ) rf(y); y x +1 rf(x ) rf(y); y x +1 = ; y = ; y x +1 x +1 ; y x +1 ; y x +1 because x +1 2 H. We ow have D f y ; x Df (x +1 ; x ) + ; y x +1 = D f (x +1 ; x ) + ; y ; x +1 D f (x +1 ; x ) + ; y + kx +1 k : Hece lm sup D f!+1 y; x lm sup D f (x +1 ; x )!+1 + lm sup!+1 ; y + lm sup!+1 kx +1 k : Sce lm!+1 = 0, lm sup!+1 ; y 0, ad lm!+1 D f (x +1 ; x ) = 0 (by (3.5)), we see that lm sup!+1 D f y ; x 0. Hece lm!+1 D f y ; x = 0. Proposto 5 ow mples that lm!+1 y x = 0. Now let xj j2n be a weakly coverget subsequece of fx g 2N ad deote ts weak lmt by o v. The y j also coverges weakly to v for ay = 1; 2; : : : ; N. Sce j2n lm f!+1 > 0 ad lm!+1 = 0, t follows from Proposto 1 that (3.6) = 1 rf(x ) rf(y ) +! 0; for ay = 1; 2; : : : ; N. Sce 2 Ay ad A s mootoe, t follows that ; z y 0

15 PROXIMAL-TYPE ALGORITHM 15 for all (z; ) 2 graph (A ). Ths, tur, mples that h; z v 0 for all (z; ) 2 graph (A ). Therefore, usg the maxmal mootocty of A, we ow obta that v 2 A 1 (0 ) for each = 1; 2; : : : ; N. Thus v 2 Z ad ths proves Clam 3. Clam 4: The sequece fx g 2N coverges strogly to proj f Z (x 0). Let ~u = proj f Z (x 0). Sce x +1 = proj f H \W (x 0 ) ad Z s cotaed H \ W, we have D f (x +1 ; x 0 ) D f (~u; x 0 ). The three pot detty (see (2.1)) yelds D f (x ; ~u) = D f (x ; x 0 ) + D f (x 0 ; ~u) hrf(~u) rf(x 0 ); x x 0 D f (~u; x 0 ) + D f (x 0 ; ~u) hrf(~u) rf(x 0 ); x x 0 = hrf(~u) rf(x 0 ); ~u x 0 hrf(~u) rf(x 0 ); x x 0 = hrf(~u) rf(x 0 ); ~u x : Now let fx g 2N be a weakly coverget subsequece of fx g 2N ad deote ts weak lmt by v. We already kow (see Clam 3) that v 2 Z. It follows from Proposto 6() that Hece lm sup D f (x ; ~u) hrf(~u) rf(x 0 ); ~u v 0:!+1 lm!+1 D f (x ; ~u) = 0: Proposto 4 ow mples that x! ~u. It follows that the whole sequece fx g 2N coverges strogly to ~u = proj f Z (x 0), as clamed. Theorem 1. Ths completes the proof of

16 16 SIMEON REICH AND SHOHAM SABACH 4. Zero Free Operators Suppose ow that the operators A, = 1; 2; : : : ; N, have o commo zero. If fx g 2N s a sequece satsfyg (3.1), the lm!+1 kx k = +1. Ths s because f fx g 2N were to have a bouded subsequece, the t would follow from Clam 3 the proof of Theorem 1 that the operators A, = 1; 2; : : : ; N, dd share a commo zero. I the case of a sgle zero free operator A, we ca prove that such a sequece always exsts. To ths ed, we rst recall the dualty mappg of the space X,.e., the mappg J : X! 2 X whch s de ed by Jx = o 2 X : h; x = kxk 2 = kk 2 : We cotue wth the followg lemma. Lemma 1. If A : X! 2 X s a maxmal mootoe operator wth a bouded e ectve doma, the A 1 (0 ) 6=?. Proof. Let f" g 2N be a sequece of postve umbers whch coverges to zero. The operator A + " J s surjectve for ay 2 N because A s a maxmal mootoe operator (see [19, Theorem 3.11, p. 166]). Therefore, for ay 2 N, there exsts x 2 dom A such that 0 2 (A + " J) x. Cosequetly, for ay 2 N, there are 2 Ax ad 2 Jx such that + " = 0. Therefore we have k k = " k k = " kx k! 0 because fx g 2N s a bouded sequece. Hece there exsts a subsequece fx k g k2n of fx g 2N whch coverges weakly to some x 0 2 X. Sce A s mootoe we have (4.1) h k ; v x k 0; k 2 N for ay (v; ) 2 graph A. Lettg k! +1 (4.1), we obta h; v x 0 0 for all (v; ) 2 graph A ad from the maxmalty of A t follows that x 0 2 A 1 (0 ). Hece A 1 (0 ) 6=?, as clamed.

17 PROXIMAL-TYPE ALGORITHM 17 Theorem 2. Let A : X! 2 X be a maxmal mootoe operator. Let f : X! R be a Legedre fucto whch s bouded, uformly Fréchet d eretable ad totally covex o bouded subsets of X. The, for each x 0 2 X, there are sequeces fx g 2N whch satsfy (3.1). If lm f!+1 > 0, ad the sequece of errors f g 2N X sats es lm!+1 = 0 ad lm sup!+1 h ; y 0, the ether A 1 (0 ) 6=? ad each such sequece fx g 2N coverges strogly to proj f A 1 (0 ) (x 0) or A 1 (0 ) =? ad each such sequece fx g 2N sats es lm!+1 kx k = +1. Proof. I vew of Theorem 1, we oly eed to cosder the case where A 1 (0 ) =?. Frst of all we prove that ths case, for each x 0 2 X, there s a sequece fx g 2N whch sats es (3.1) wth = 0 for all 2 N. We prove ths by ducto. We rst check that the tal step ( = 0) s well de ed. The problem 0 2 Ax (rf(x) rf(x 0 )) always has a soluto (y 0 ; 0 ) because t s equvalet to the problem x =Res f (x 0A 0) ad ths problem does have a soluto sce dom Res f A = X (see Proposto 3 ad [5, Theorem 3.13(v), p. 606]). Now ote that W 0 = X. Sce H 0 caot be empty, the ext terate x 1 ca be geerated; t s the Bregma projecto of x 0 oto H 0 = W 0 \ H 0. Note that wheever x s geerated, y ad ca further be obtaed because the proxmal subproblems always have solutos. Suppose ow that x ad (y ; ) have already bee de ed for = 0; : : : ; ^. We have to prove that x^+1 s also well de ed. To ths ed, take ay z 0 2 dom A ad de e = max fky z 0 k : = 0; : : : ; ^g ad 8 >< 0; kx z 0 k + 1 h(x) = >: +1; otherwse:

18 18 SIMEON REICH AND SHOHAM SABACH The h : X! ( 1; +1] s a proper, covex ad lower semcotuous fucto, ts subd s maxmal mootoe (see [28, Theorem 2.13, p. 124]), ad A 0 = A s also maxmal mootoe (see [33]). Furthermore, A 0 (z) = A (z) for all kz z 0 k < + 1: Therefore 2 A 0 y for = 0; : : : ; ^. We coclude that x ad (y ; ) also satsfy the codtos of Theorem 1 appled to the problem 0 2 A 0 (x). Sce A 0 has a bouded e ectve doma, ths problem has a soluto by Lemma 1. Thus t follows from Clam 1 the proof of Theorem 1 that x^+1 s well de ed. Hece the whole sequece fx g 2N s well de ed, as asserted. If fx g 2N were to have a bouded subsequece, the t would follow from Clam 3 the proof of Theorem 1 that A had a zero. Therefore f A 1 (0 ) =?, the lm!+1 kx k = +1, as asserted. Remark 1. I both Theorems 1 ad 2 we ca replace the assumptos that lm f!+1 > 0 ad f s uformly Fréchet d eretable o bouded subsets of X wth the assumpto that lm!+1 = +1. Ths s because ths case frf(x ) rf(y )g 2N s bouded ad therefore (3.6) cotues to hold. 5. Cosequeces of the Strog Covergece Theorem Algorthm (1.4) s a specal case of algorthm (3.1) whe N = 1 ad = 0 for all 2 N. Hece as a drect cosequece of Theorem 1 we obta the followg result (cf. [21]). Corollary 1. Let A : X! 2 X be a maxmal mootoe operator. Let f : X! R be a Legedre fucto whch s bouded, uformly Fréchet d eretable ad totally covex o bouded subsets of X, ad suppose that lm f!+1 > 0. The for each x 0 2 X, the sequece fx g 2N geerated by (1.4) s well de ed,

19 PROXIMAL-TYPE ALGORITHM 19 ad ether A 1 (0 ) 6=? ad fx g 2N coverges strogly to proj f A 1 (0 ) (x 0) as! +1, or A 1 (0 ) =? ad lm!+1 kx k = +1. Notable corollares of Theorems 1 ad 2 occur whe the space X s both uformly smooth ad uformly covex. I ths case the fucto f(x) = 1 2 kxk2 s Legedre (cf. [3, Lemma 6.2, p.24]) ad uformly Fréchet d eretable o bouded subsets of X. Accordg to [13, Corollary 1(), p. 325], f s sequetally cosstet sce X s uformly covex ad hece f s totally covex o bouded subsets of X. Therefore Theorems 1 ad 2 hold ths cotext ad lead us to the followg two results whch, some sese, complemet Theorem 8 [25] (see also Theorem 1 [37]). Corollary 2. Let X be a uformly smooth ad uformly covex Baach space ad let A : X! 2 X be a maxmal mootoe operator. The, for each x 0 2 X, the sequece fx g 2N geerated by (1.3) s well de ed. If lm f!+1 > 0, the ether A 1 (0 ) 6=? ad fx g 2N coverges strogly to proj f A 1 (0 ) (x 0) as! +1, or A 1 (0 ) =? ad lm!+1 kx k = +1. Corollary 3. Let X be a Hlbert space ad let A : X! 2 X be a maxmal mootoe operator. The, for each x 0 2 X, the sequece fx g 2N geerated by (1.2) s well de ed. If lm f!1 > 0, the ether A 1 (0) 6=? ad fx g 2N coverges strogly to P A 1 (0)(x 0 ) as! +1, or A 1 (0) =? ad lm!+1 kx k = +1. These corollares also hold, of course, the presece of computatoal errors as Theorems 1 ad A Applcato of the Strog Covergece Theorem Let g : X! ( 1; +1] be a proper, covex ad lower semcotuous fucto. Recall that the subd of g s de ed for ay x 2 X (x) := f 2 X : h; y x g (y) g (x) 8y 2 Xg :

20 20 SIMEON REICH AND SHOHAM SABACH Usg Theorem 1 ad the subd eretal of g, we obta a algorthm for dg a mmzer of g. Proposto 7. Let g : X! ( 1; +1] be a proper, covex ad lower semcotuous fucto whch attas ts mmum over X. If f : X! R s a Legedre fucto whch s bouded, uformly Fréchet d eretable, ad totally covex o bouded subsets of X, ad f g 2N s a postve sequece wth lm f!+1 > 0, the, for each x 0 2 X, the sequece fx g 2N geerated by 8 x 0 2 X; >< 0 = + 1 (rf(y ) rf(x )) ; (y ) ; H = fz 2 X : h ; z y 0g ; W = fz 2 X : hrf(x 0 ) rf(x ); z x 0g ; >: x +1 = proj f H \W (x 0 ); = 0; 1; 2; : : : ; coverges strogly to a mmzer of g as! +1. If g does ot atta ts mmum over X, the lm!+1 kx k = +1. Proof. The subd of g s a maxmal mootoe operator sce g s a proper, covex ad lower semcotuous fucto (see [28, Theorem 2.13, p. 124]). Sce the zero set cocdes wth the set of mmzers of g, Proposto 7 follows mmedately from Theorems 1 ad 2. Note that ths case y = arg m g (x) + 1 D f (x; x ) x2x s equvalet to 0 (y ) + 1 (rf(y ) rf(x )) :

21 PROXIMAL-TYPE ALGORITHM Ackowledgemet The rst author was partally supported by the Israel Scece Foudato (Grat 647/07), by the Fud for the Promoto of Research at the Techo, ad by the Techo Presdet s Research Fud. Refereces [1] Ambrosett, A. ad Prod, G.: A prmer of olear aalyss, Cambrdge Uversty Press, Cambrdge, [2] Bauschke, H. H. ad Borwe, J. M.: Legedre fuctos ad the method of radom Bregma projectos, J. Covex Aal. 4 (1997), [3] Bauschke, H. H., Borwe, J. M. ad Combettes, P. L.: Essetal smoothess, essetal strct covexty, ad Legedre fuctos Baach spaces, Commu. Cotemp. Math. 3 (2001), [4] Bauschke, H. H. ad Combettes, P. L.: A weak-to-strog covergece prcple for Fejérmootoe methods Hlbert spaces, Math. Oper. Res. 26 (2001), [5] Bauschke, H. H., Borwe, J. M. ad Combettes, P. L.: Bregma mootoe optmzato algorthms, SIAM J. Cotrol Optm. 42 (2003), [6] Bauschke, H. H., Matoušková, E. ad Rech, S.: Projecto ad proxmal pot methods: covergece results ad couterexamples, Nolear Aal. 56 (2004), [7] Boas, J. F. ad Shapro, A.: Perturbato aalyss of optmzato problems, Sprger Verlag, New York, [8] Bregma, L. M.: A relaxato method for dg the commo pot of covex sets ad ts applcato to the soluto of problems covex programmg, USSR Comput. Math. ad Math. Phys. 7 (1967), [9] Brézs, H. ad Los, P.-L.: Produts s de résolvates, Israel J. Math. 29 (1978), [10] Bruck, R. E. ad Rech, S.: Noexpasve projectos ad resolvets of accretve operators, Housto J. Math. 3 (1977), [11] Butaru, D., Cesor, Y. ad Rech, S.: Iteratve averagg of etropc projectos for solvg stochastc covex feasblty problems, Computatoal Optmzato ad Applcatos 8 (1997), [12] Butaru, D. ad Iusem, A. N.: Totally covex fuctos for xed pots computato ad te dmesoal optmzato, Kluwer Academc Publshers, Dordrecht, 2000.

22 22 SIMEON REICH AND SHOHAM SABACH [13] Butaru, D., Iusem, A. N. ad Resmerta, E.: Total covexty for powers of the orm uformly covex Baach spaces, J. Covex Aal. 7 (2000), [14] Butaru, D., Iusem, A. N. ad Z¼alescu, C.: O uform covexty, total covexty ad covergece of the proxmal pot ad outer Bregma projecto algorthms Baach spaces, J. Covex Aal. 10 (2003), [15] Butaru, D. ad Resmerta, E.: Bregma dstaces, totally covex fuctos ad a method for solvg operator equatos Baach spaces, Abstr. Appl. Aal. 2006, Art. ID 84919, [16] Butaru, D. ad Kassay, G.: A proxmal-projecto method for dg zeroes of set-valued operators, SIAM J. Cotrol Optm. 47 (2008), [17] Cesor, Y. ad Let, A.: A teratve row-acto method for terval covex programmg, J. Optm. Theory Appl. 34 (1981), [18] Cesor, Y. ad Zeos, S. A.: Proxmal mmzato algorthm wth D-fuctos, J. Optm. Theory Appl. 73 (1992), [19] Coraescu, I.: Geometry of Baach spaces, dualty mappgs ad olear problems, Kluwer Academc Publshers, Dordrecht, [20] Eckste, J.: Nolear proxmal pot algorthms usg Bregma fuctos, wth applcato to covex programmg, Math. Oper. Res. 18 (1993), [21] Gárcga Otero, R. ad Svater, B. F.: A strogly coverget hybrd proxmal method Baach spaces, J. Math. Aal. Appl. 289 (2004), [22] Güler, O.: O the covergece of the proxmal pot algorthm for covex mmzato, SIAM J. Cotrol Optm. 29 (1991), [23] Kammura, S. ad Takahash, W.: Approxmatg solutos of maxmal mootoe operators Hlbert spaces, J. Approx. Theory 106 (2000), [24] Kammura, S. ad Takahash, W.: Weak ad strog covergece of solutos to accretve operator clusos ad applcatos, Set-Valued Aal. 8 (2000), [25] Kammura, S. ad Takahash, W.: Strog covergece of a proxmal-type algorthm a Baach space, SIAM J. Optm. 13 (2002), [26] Martet, B.: Régularsato d équatos varatoelles par approxmatos successves, Revue Fraçase d Iformatque et de Recherche Opératoelle 4 (1970), [27] Nevala, O. ad Rech, S.: Strog covergece of cotracto semgroups ad of teratve methods for accretve operators Baach spaces, Israel J. Math. 32 (1979), [28] Pascal, D. ad Sburla, S.: Nolear mappgs of mootoe type, Sjtho & Nordho Iteratoal Publshers, Alphe aa de Rj, 1978.

23 PROXIMAL-TYPE ALGORITHM 23 [29] Rech, S.: Weak covergece theorems for oexpasve mappgs Baach spaces, J. Math. Aal. Appl. 67 (1979), [30] Rech, S.: A weak covergece theorem for the alteratg method wth Bregma dstaces, Theory ad applcatos of olear operators of accretve ad mootoe type, Marcel Dekker, New York, 1996, [31] Rech, S. ad Zaslavsk, A. J.: I te products of resolvets of accretve operator, Topol. Methods Nolear Aal. 15 (2000), [32] Resmerta, E.: O total covexty, Bregma projectos ad stablty Baach spaces, J. Covex Aal. 11 (2004), [33] Rockafellar, R. T.: O the maxmalty of sums of olear mootoe operators, Tras. Amer. Math. Soc. 149 (1970), [34] Rockafellar, R. T.: Mootoe operators ad the proxmal pot algorthm, SIAM J. Cotrol Optm. 14 (1976), [35] Rockafellar, R. T.: Augmeted Lagragas ad applcatos of the proxmal pot algorthm covex programmg, Math. Oper. Res. 1 (1976), [36] Rockafellar, R. T. ad Wets, R. J.-B.: Varatoal aalyss, Sprger Verlag, Berl, [37] Solodov, M. V. ad Svater, B. F.: Forcg strog covergece of proxmal pot teratos a Hlbert space, Math. Program. 87 (2000), [38] Yao, J. C. ad Zeg, L. C.: A exact proxmal-type algorthm Baach spaces, J. Optm. Theory Appl. 135 (2007), [39] Z¼alescu, C.: Covex aalyss geeral vector spaces, World Scet c Publshg, Sgapore, Smeo Rech: Departmet of Mathematcs, The Techo Israel Isttute of Techology, Hafa, Israel E-mal address: srech@tx.techo.ac.l Shoham Sabach: Departmet of Mathematcs, The Techo Israel Isttute of Techology, Hafa, Israel E-mal address: ssabach@tx.techo.ac.l

Strong Convergence of Weighted Averaged Approximants of Asymptotically Nonexpansive Mappings in Banach Spaces without Uniform Convexity

Strong Convergence of Weighted Averaged Approximants of Asymptotically Nonexpansive Mappings in Banach Spaces without Uniform Convexity BULLETIN of the MALAYSIAN MATHEMATICAL SCIENCES SOCIETY Bull. Malays. Math. Sc. Soc. () 7 (004), 5 35 Strog Covergece of Weghted Averaged Appromats of Asymptotcally Noepasve Mappgs Baach Spaces wthout

More information

Research Article A New Iterative Method for Common Fixed Points of a Finite Family of Nonexpansive Mappings

Research Article A New Iterative Method for Common Fixed Points of a Finite Family of Nonexpansive Mappings Hdaw Publshg Corporato Iteratoal Joural of Mathematcs ad Mathematcal Sceces Volume 009, Artcle ID 391839, 9 pages do:10.1155/009/391839 Research Artcle A New Iteratve Method for Commo Fxed Pots of a Fte

More information

PROJECTION PROBLEM FOR REGULAR POLYGONS

PROJECTION PROBLEM FOR REGULAR POLYGONS Joural of Mathematcal Sceces: Advaces ad Applcatos Volume, Number, 008, Pages 95-50 PROJECTION PROBLEM FOR REGULAR POLYGONS College of Scece Bejg Forestry Uversty Bejg 0008 P. R. Cha e-mal: sl@bjfu.edu.c

More information

Complete Convergence and Some Maximal Inequalities for Weighted Sums of Random Variables

Complete Convergence and Some Maximal Inequalities for Weighted Sums of Random Variables Joural of Sceces, Islamc Republc of Ira 8(4): -6 (007) Uversty of Tehra, ISSN 06-04 http://sceces.ut.ac.r Complete Covergece ad Some Maxmal Iequaltes for Weghted Sums of Radom Varables M. Am,,* H.R. Nl

More information

THE PROBABILISTIC STABILITY FOR THE GAMMA FUNCTIONAL EQUATION

THE PROBABILISTIC STABILITY FOR THE GAMMA FUNCTIONAL EQUATION Joural of Scece ad Arts Year 12, No. 3(2), pp. 297-32, 212 ORIGINAL AER THE ROBABILISTIC STABILITY FOR THE GAMMA FUNCTIONAL EQUATION DOREL MIHET 1, CLAUDIA ZAHARIA 1 Mauscrpt receved: 3.6.212; Accepted

More information

A Remark on the Uniform Convergence of Some Sequences of Functions

A Remark on the Uniform Convergence of Some Sequences of Functions Advaces Pure Mathematcs 05 5 57-533 Publshed Ole July 05 ScRes. http://www.scrp.org/joural/apm http://dx.do.org/0.436/apm.05.59048 A Remark o the Uform Covergece of Some Sequeces of Fuctos Guy Degla Isttut

More information

18.413: Error Correcting Codes Lab March 2, Lecture 8

18.413: Error Correcting Codes Lab March 2, Lecture 8 18.413: Error Correctg Codes Lab March 2, 2004 Lecturer: Dael A. Spelma Lecture 8 8.1 Vector Spaces A set C {0, 1} s a vector space f for x all C ad y C, x + y C, where we take addto to be compoet wse

More information

= lim. (x 1 x 2... x n ) 1 n. = log. x i. = M, n

= lim. (x 1 x 2... x n ) 1 n. = log. x i. = M, n .. Soluto of Problem. M s obvously cotuous o ], [ ad ], [. Observe that M x,..., x ) M x,..., x ) )..) We ext show that M s odecreasg o ], [. Of course.) mles that M s odecreasg o ], [ as well. To show

More information

The Mathematical Appendix

The Mathematical Appendix The Mathematcal Appedx Defto A: If ( Λ, Ω, where ( λ λ λ whch the probablty dstrbutos,,..., Defto A. uppose that ( Λ,,..., s a expermet type, the σ-algebra o λ λ λ are defed s deoted by ( (,,...,, σ Ω.

More information

Non-uniform Turán-type problems

Non-uniform Turán-type problems Joural of Combatoral Theory, Seres A 111 2005 106 110 wwwelsevercomlocatecta No-uform Turá-type problems DhruvMubay 1, Y Zhao 2 Departmet of Mathematcs, Statstcs, ad Computer Scece, Uversty of Illos at

More information

Aitken delta-squared generalized Juncgk-type iterative procedure

Aitken delta-squared generalized Juncgk-type iterative procedure Atke delta-squared geeralzed Jucgk-type teratve procedure M. De la Se Isttute of Research ad Developmet of Processes. Uversty of Basque Coutry Campus of Leoa (Bzkaa) PO Box. 644- Blbao, 488- Blbao. SPAIN

More information

MATH 247/Winter Notes on the adjoint and on normal operators.

MATH 247/Winter Notes on the adjoint and on normal operators. MATH 47/Wter 00 Notes o the adjot ad o ormal operators I these otes, V s a fte dmesoal er product space over, wth gve er * product uv, T, S, T, are lear operators o V U, W are subspaces of V Whe we say

More information

Entropy ISSN by MDPI

Entropy ISSN by MDPI Etropy 2003, 5, 233-238 Etropy ISSN 1099-4300 2003 by MDPI www.mdp.org/etropy O the Measure Etropy of Addtve Cellular Automata Hasa Aı Arts ad Sceces Faculty, Departmet of Mathematcs, Harra Uversty; 63100,

More information

International Journal of Mathematical Archive-5(8), 2014, Available online through ISSN

International Journal of Mathematical Archive-5(8), 2014, Available online through   ISSN Iteratoal Joural of Mathematcal Archve-5(8) 204 25-29 Avalable ole through www.jma.fo ISSN 2229 5046 COMMON FIXED POINT OF GENERALIZED CONTRACTION MAPPING IN FUZZY METRIC SPACES Hamd Mottagh Golsha* ad

More information

On the convergence of derivatives of Bernstein approximation

On the convergence of derivatives of Bernstein approximation O the covergece of dervatves of Berste approxmato Mchael S. Floater Abstract: By dfferetatg a remader formula of Stacu, we derve both a error boud ad a asymptotc formula for the dervatves of Berste approxmato.

More information

Lebesgue Measure of Generalized Cantor Set

Lebesgue Measure of Generalized Cantor Set Aals of Pure ad Appled Mathematcs Vol., No.,, -8 ISSN: -8X P), -888ole) Publshed o 8 May www.researchmathsc.org Aals of Lebesgue Measure of Geeralzed ator Set Md. Jahurul Islam ad Md. Shahdul Islam Departmet

More information

Q-analogue of a Linear Transformation Preserving Log-concavity

Q-analogue of a Linear Transformation Preserving Log-concavity Iteratoal Joural of Algebra, Vol. 1, 2007, o. 2, 87-94 Q-aalogue of a Lear Trasformato Preservg Log-cocavty Daozhog Luo Departmet of Mathematcs, Huaqao Uversty Quazhou, Fua 362021, P. R. Cha ldzblue@163.com

More information

. The set of these sums. be a partition of [ ab, ]. Consider the sum f( x) f( x 1)

. The set of these sums. be a partition of [ ab, ]. Consider the sum f( x) f( x 1) Chapter 7 Fuctos o Bouded Varato. Subject: Real Aalyss Level: M.Sc. Source: Syed Gul Shah (Charma, Departmet o Mathematcs, US Sargodha Collected & Composed by: Atq ur Rehma (atq@mathcty.org, http://www.mathcty.org

More information

Variable Metric Forward-Backward Splitting with Applications to Monotone Inclusions in Duality

Variable Metric Forward-Backward Splitting with Applications to Monotone Inclusions in Duality Varable Metrc Forward-Backward Splttg wth Applcatos to Mootoe Iclusos Dualty Patrck L. Combettes ad B`ăg Côg Vũ UPMC Uversté Pars 06 Laboratore Jacques-Lous Los UMR CNRS 7598 75005 Pars, Frace plc@math.jusseu.fr,

More information

Uniform asymptotical stability of almost periodic solution of a discrete multispecies Lotka-Volterra competition system

Uniform asymptotical stability of almost periodic solution of a discrete multispecies Lotka-Volterra competition system Iteratoal Joural of Egeerg ad Advaced Research Techology (IJEART) ISSN: 2454-9290, Volume-2, Issue-1, Jauary 2016 Uform asymptotcal stablty of almost perodc soluto of a dscrete multspeces Lotka-Volterra

More information

1 0, x? x x. 1 Root finding. 1.1 Introduction. Solve[x^2-1 0,x] {{x -1},{x 1}} Plot[x^2-1,{x,-2,2}] 3

1 0, x? x x. 1 Root finding. 1.1 Introduction. Solve[x^2-1 0,x] {{x -1},{x 1}} Plot[x^2-1,{x,-2,2}] 3 Adrew Powuk - http://www.powuk.com- Math 49 (Numercal Aalyss) Root fdg. Itroducto f ( ),?,? Solve[^-,] {{-},{}} Plot[^-,{,-,}] Cubc equato https://e.wkpeda.org/wk/cubc_fucto Quartc equato https://e.wkpeda.org/wk/quartc_fucto

More information

Maps on Triangular Matrix Algebras

Maps on Triangular Matrix Algebras Maps o ragular Matrx lgebras HMED RMZI SOUROUR Departmet of Mathematcs ad Statstcs Uversty of Vctora Vctora, BC V8W 3P4 CND sourour@mathuvcca bstract We surveys results about somorphsms, Jorda somorphsms,

More information

INTEGRATION THEORY AND FUNCTIONAL ANALYSIS MM-501

INTEGRATION THEORY AND FUNCTIONAL ANALYSIS MM-501 INTEGRATION THEORY AND FUNCTIONAL ANALYSIS M.A./M.Sc. Mathematcs (Fal) MM-50 Drectorate of Dstace Educato Maharsh Dayaad Uversty ROHTAK 4 00 Copyrght 004, Maharsh Dayaad Uversty, ROHTAK All Rghts Reserved.

More information

CHAPTER 3 COMMON FIXED POINT THEOREMS IN GENERALIZED SPACES

CHAPTER 3 COMMON FIXED POINT THEOREMS IN GENERALIZED SPACES CHAPTER 3 COMMON FIXED POINT THEOREMS IN GENERAIZED SPACES 3. INTRODUCTION After the celebrated Baach cotracto prcple (BCP) 9, there have bee umerous results the lterature dealg wth mappgs satsfyg the

More information

arxiv: v1 [cs.lg] 22 Feb 2015

arxiv: v1 [cs.lg] 22 Feb 2015 SDCA wthout Dualty Sha Shalev-Shwartz arxv:50.0677v cs.lg Feb 05 Abstract Stochastc Dual Coordate Ascet s a popular method for solvg regularzed loss mmzato for the case of covex losses. I ths paper we

More information

Abstract. 1. Introduction

Abstract. 1. Introduction Joura of Mathematca Sceces: Advaces ad Appcatos Voume 4 umber 2 2 Pages 33-34 COVERGECE OF HE PROJECO YPE SHKAWA ERAO PROCESS WH ERRORS FOR A FE FAMY OF OSEF -ASYMPOCAY QUAS-OEXPASVE MAPPGS HUA QU ad S-SHEG

More information

Generalized Convex Functions on Fractal Sets and Two Related Inequalities

Generalized Convex Functions on Fractal Sets and Two Related Inequalities Geeralzed Covex Fuctos o Fractal Sets ad Two Related Iequaltes Huxa Mo, X Su ad Dogya Yu 3,,3School of Scece, Bejg Uversty of Posts ad Telecommucatos, Bejg,00876, Cha, Correspodece should be addressed

More information

On the Subdifferentials of Quasiconvex and Pseudoconvex Functions and Cyclic Monotonicity*

On the Subdifferentials of Quasiconvex and Pseudoconvex Functions and Cyclic Monotonicity* Ž. Joural of Mathematcal Aalyss ad Applcatos 237, 3042 1999 Artcle ID jmaa.1999.6437, avalable ole at http:www.dealbrary.com o O the Subdfferetals of Quascovex ad Pseudocovex Fuctos ad Cyclc Mootocty*

More information

1 Onto functions and bijections Applications to Counting

1 Onto functions and bijections Applications to Counting 1 Oto fuctos ad bectos Applcatos to Coutg Now we move o to a ew topc. Defto 1.1 (Surecto. A fucto f : A B s sad to be surectve or oto f for each b B there s some a A so that f(a B. What are examples of

More information

Solution of General Dual Fuzzy Linear Systems. Using ABS Algorithm

Solution of General Dual Fuzzy Linear Systems. Using ABS Algorithm Appled Mathematcal Sceces, Vol 6, 0, o 4, 63-7 Soluto of Geeral Dual Fuzzy Lear Systems Usg ABS Algorthm M A Farborz Aragh * ad M M ossezadeh Departmet of Mathematcs, Islamc Azad Uversty Cetral ehra Brach,

More information

arxiv:math/ v1 [math.gm] 8 Dec 2005

arxiv:math/ v1 [math.gm] 8 Dec 2005 arxv:math/05272v [math.gm] 8 Dec 2005 A GENERALIZATION OF AN INEQUALITY FROM IMO 2005 NIKOLAI NIKOLOV The preset paper was spred by the thrd problem from the IMO 2005. A specal award was gve to Yure Boreko

More information

Assignment 5/MATH 247/Winter Due: Friday, February 19 in class (!) (answers will be posted right after class)

Assignment 5/MATH 247/Winter Due: Friday, February 19 in class (!) (answers will be posted right after class) Assgmet 5/MATH 7/Wter 00 Due: Frday, February 9 class (!) (aswers wll be posted rght after class) As usual, there are peces of text, before the questos [], [], themselves. Recall: For the quadratc form

More information

Mu Sequences/Series Solutions National Convention 2014

Mu Sequences/Series Solutions National Convention 2014 Mu Sequeces/Seres Solutos Natoal Coveto 04 C 6 E A 6C A 6 B B 7 A D 7 D C 7 A B 8 A B 8 A C 8 E 4 B 9 B 4 E 9 B 4 C 9 E C 0 A A 0 D B 0 C C Usg basc propertes of arthmetc sequeces, we fd a ad bm m We eed

More information

The Arithmetic-Geometric mean inequality in an external formula. Yuki Seo. October 23, 2012

The Arithmetic-Geometric mean inequality in an external formula. Yuki Seo. October 23, 2012 Sc. Math. Japocae Vol. 00, No. 0 0000, 000 000 1 The Arthmetc-Geometrc mea equalty a exteral formula Yuk Seo October 23, 2012 Abstract. The classcal Jese equalty ad ts reverse are dscussed by meas of terally

More information

Journal of Mathematical Analysis and Applications

Journal of Mathematical Analysis and Applications J. Math. Aal. Appl. 365 200) 358 362 Cotets lsts avalable at SceceDrect Joural of Mathematcal Aalyss ad Applcatos www.elsever.com/locate/maa Asymptotc behavor of termedate pots the dfferetal mea value

More information

On Modified Interval Symmetric Single-Step Procedure ISS2-5D for the Simultaneous Inclusion of Polynomial Zeros

On Modified Interval Symmetric Single-Step Procedure ISS2-5D for the Simultaneous Inclusion of Polynomial Zeros It. Joural of Math. Aalyss, Vol. 7, 2013, o. 20, 983-988 HIKARI Ltd, www.m-hkar.com O Modfed Iterval Symmetrc Sgle-Step Procedure ISS2-5D for the Smultaeous Icluso of Polyomal Zeros 1 Nora Jamalud, 1 Masor

More information

Complete Convergence for Weighted Sums of Arrays of Rowwise Asymptotically Almost Negative Associated Random Variables

Complete Convergence for Weighted Sums of Arrays of Rowwise Asymptotically Almost Negative Associated Random Variables A^VÇÚO 1 32 ò 1 5 Ï 2016 c 10 Chese Joural of Appled Probablty ad Statstcs Oct., 2016, Vol. 32, No. 5, pp. 489-498 do: 10.3969/j.ss.1001-4268.2016.05.005 Complete Covergece for Weghted Sums of Arrays of

More information

1 Lyapunov Stability Theory

1 Lyapunov Stability Theory Lyapuov Stablty heory I ths secto we cosder proofs of stablty of equlbra of autoomous systems. hs s stadard theory for olear systems, ad oe of the most mportat tools the aalyss of olear systems. It may

More information

Two Strong Convergence Theorems for a Proximal Method in Reflexive Banach Spaces

Two Strong Convergence Theorems for a Proximal Method in Reflexive Banach Spaces Two Strong Convergence Theorems for a Proxmal Method n Reflexve Banach Spaces Smeon Rech and Shoham Sabach Abstract. Two strong convergence theorems for a proxmal method for fndng common zeroes of maxmal

More information

BERNSTEIN COLLOCATION METHOD FOR SOLVING NONLINEAR DIFFERENTIAL EQUATIONS. Aysegul Akyuz Dascioglu and Nese Isler

BERNSTEIN COLLOCATION METHOD FOR SOLVING NONLINEAR DIFFERENTIAL EQUATIONS. Aysegul Akyuz Dascioglu and Nese Isler Mathematcal ad Computatoal Applcatos, Vol. 8, No. 3, pp. 293-300, 203 BERNSTEIN COLLOCATION METHOD FOR SOLVING NONLINEAR DIFFERENTIAL EQUATIONS Aysegul Ayuz Dascoglu ad Nese Isler Departmet of Mathematcs,

More information

Generalization of the Dissimilarity Measure of Fuzzy Sets

Generalization of the Dissimilarity Measure of Fuzzy Sets Iteratoal Mathematcal Forum 2 2007 o. 68 3395-3400 Geeralzato of the Dssmlarty Measure of Fuzzy Sets Faramarz Faghh Boformatcs Laboratory Naobotechology Research Ceter vesa Research Isttute CECR Tehra

More information

Part 4b Asymptotic Results for MRR2 using PRESS. Recall that the PRESS statistic is a special type of cross validation procedure (see Allen (1971))

Part 4b Asymptotic Results for MRR2 using PRESS. Recall that the PRESS statistic is a special type of cross validation procedure (see Allen (1971)) art 4b Asymptotc Results for MRR usg RESS Recall that the RESS statstc s a specal type of cross valdato procedure (see Alle (97)) partcular to the regresso problem ad volves fdg Y $,, the estmate at the

More information

LECTURE 24 LECTURE OUTLINE

LECTURE 24 LECTURE OUTLINE LECTURE 24 LECTURE OUTLINE Gradet proxmal mmzato method Noquadratc proxmal algorthms Etropy mmzato algorthm Expoetal augmeted Lagraga mehod Etropc descet algorthm **************************************

More information

PTAS for Bin-Packing

PTAS for Bin-Packing CS 663: Patter Matchg Algorthms Scrbe: Che Jag /9/00. Itroducto PTAS for B-Packg The B-Packg problem s NP-hard. If we use approxmato algorthms, the B-Packg problem could be solved polyomal tme. For example,

More information

A Penalty Function Algorithm with Objective Parameters and Constraint Penalty Parameter for Multi-Objective Programming

A Penalty Function Algorithm with Objective Parameters and Constraint Penalty Parameter for Multi-Objective Programming Aerca Joural of Operatos Research, 4, 4, 33-339 Publshed Ole Noveber 4 ScRes http://wwwscrporg/oural/aor http://ddoorg/436/aor4463 A Pealty Fucto Algorth wth Obectve Paraeters ad Costrat Pealty Paraeter

More information

Functions of Random Variables

Functions of Random Variables Fuctos of Radom Varables Chapter Fve Fuctos of Radom Varables 5. Itroducto A geeral egeerg aalyss model s show Fg. 5.. The model output (respose) cotas the performaces of a system or product, such as weght,

More information

02/15/04 INTERESTING FINITE AND INFINITE PRODUCTS FROM SIMPLE ALGEBRAIC IDENTITIES

02/15/04 INTERESTING FINITE AND INFINITE PRODUCTS FROM SIMPLE ALGEBRAIC IDENTITIES 0/5/04 ITERESTIG FIITE AD IFIITE PRODUCTS FROM SIMPLE ALGEBRAIC IDETITIES Thomas J Osler Mathematcs Departmet Rowa Uversty Glassboro J 0808 Osler@rowaedu Itroducto The dfferece of two squares, y = + y

More information

arxiv: v4 [math.nt] 14 Aug 2015

arxiv: v4 [math.nt] 14 Aug 2015 arxv:52.799v4 [math.nt] 4 Aug 25 O the propertes of terated bomal trasforms for the Padova ad Perr matrx sequeces Nazmye Ylmaz ad Necat Tasara Departmet of Mathematcs, Faculty of Scece, Selcu Uversty,

More information

Discrete Mathematics and Probability Theory Fall 2016 Seshia and Walrand DIS 10b

Discrete Mathematics and Probability Theory Fall 2016 Seshia and Walrand DIS 10b CS 70 Dscrete Mathematcs ad Probablty Theory Fall 206 Sesha ad Walrad DIS 0b. Wll I Get My Package? Seaky delvery guy of some compay s out delverg packages to customers. Not oly does he had a radom package

More information

Chapter 5 Properties of a Random Sample

Chapter 5 Properties of a Random Sample Lecture 6 o BST 63: Statstcal Theory I Ku Zhag, /0/008 Revew for the prevous lecture Cocepts: t-dstrbuto, F-dstrbuto Theorems: Dstrbutos of sample mea ad sample varace, relatoshp betwee sample mea ad sample

More information

Lecture 9: Tolerant Testing

Lecture 9: Tolerant Testing Lecture 9: Tolerat Testg Dael Kae Scrbe: Sakeerth Rao Aprl 4, 07 Abstract I ths lecture we prove a quas lear lower boud o the umber of samples eeded to do tolerat testg for L dstace. Tolerat Testg We have

More information

About k-perfect numbers

About k-perfect numbers DOI: 0.47/auom-04-0005 A. Şt. Uv. Ovdus Costaţa Vol.,04, 45 50 About k-perfect umbers Mhály Becze Abstract ABSTRACT. I ths paper we preset some results about k-perfect umbers, ad geeralze two equaltes

More information

D KL (P Q) := p i ln p i q i

D KL (P Q) := p i ln p i q i Cheroff-Bouds 1 The Geeral Boud Let P 1,, m ) ad Q q 1,, q m ) be two dstrbutos o m elemets, e,, q 0, for 1,, m, ad m 1 m 1 q 1 The Kullback-Lebler dvergece or relatve etroy of P ad Q s defed as m D KL

More information

MA 524 Homework 6 Solutions

MA 524 Homework 6 Solutions MA 524 Homework 6 Solutos. Sce S(, s the umber of ways to partto [] to k oempty blocks, ad c(, s the umber of ways to partto to k oempty blocks ad also the arrage each block to a cycle, we must have S(,

More information

Extreme Value Theory: An Introduction

Extreme Value Theory: An Introduction (correcto d Extreme Value Theory: A Itroducto by Laures de Haa ad Aa Ferrera Wth ths webpage the authors ted to form the readers of errors or mstakes foud the book after publcato. We also gve extesos for

More information

X ε ) = 0, or equivalently, lim

X ε ) = 0, or equivalently, lim Revew for the prevous lecture Cocepts: order statstcs Theorems: Dstrbutos of order statstcs Examples: How to get the dstrbuto of order statstcs Chapter 5 Propertes of a Radom Sample Secto 55 Covergece

More information

F. Inequalities. HKAL Pure Mathematics. 進佳數學團隊 Dr. Herbert Lam 林康榮博士. [Solution] Example Basic properties

F. Inequalities. HKAL Pure Mathematics. 進佳數學團隊 Dr. Herbert Lam 林康榮博士. [Solution] Example Basic properties 進佳數學團隊 Dr. Herbert Lam 林康榮博士 HKAL Pure Mathematcs F. Ieualtes. Basc propertes Theorem Let a, b, c be real umbers. () If a b ad b c, the a c. () If a b ad c 0, the ac bc, but f a b ad c 0, the ac bc. Theorem

More information

ANALYSIS ON THE NATURE OF THE BASIC EQUATIONS IN SYNERGETIC INTER-REPRESENTATION NETWORK

ANALYSIS ON THE NATURE OF THE BASIC EQUATIONS IN SYNERGETIC INTER-REPRESENTATION NETWORK Far East Joural of Appled Mathematcs Volume, Number, 2008, Pages Ths paper s avalable ole at http://www.pphm.com 2008 Pushpa Publshg House ANALYSIS ON THE NATURE OF THE ASI EQUATIONS IN SYNERGETI INTER-REPRESENTATION

More information

A nonsmooth Levenberg-Marquardt method for generalized complementarity problem

A nonsmooth Levenberg-Marquardt method for generalized complementarity problem ISSN 746-7659 Egla UK Joural of Iformato a Computg Scece Vol. 7 No. 4 0 pp. 67-7 A osmooth Leveberg-Marquart metho for geeralze complemetarty problem Shou-qag Du College of Mathematcs Qgao Uversty Qgao

More information

{ }{ ( )} (, ) = ( ) ( ) ( ) Chapter 14 Exercises in Sampling Theory. Exercise 1 (Simple random sampling): Solution:

{ }{ ( )} (, ) = ( ) ( ) ( ) Chapter 14 Exercises in Sampling Theory. Exercise 1 (Simple random sampling): Solution: Chapter 4 Exercses Samplg Theory Exercse (Smple radom samplg: Let there be two correlated radom varables X ad A sample of sze s draw from a populato by smple radom samplg wthout replacemet The observed

More information

Lecture 16: Backpropogation Algorithm Neural Networks with smooth activation functions

Lecture 16: Backpropogation Algorithm Neural Networks with smooth activation functions CO-511: Learg Theory prg 2017 Lecturer: Ro Lv Lecture 16: Bacpropogato Algorthm Dsclamer: These otes have ot bee subected to the usual scruty reserved for formal publcatos. They may be dstrbuted outsde

More information

v 1 -periodic 2-exponents of SU(2 e ) and SU(2 e + 1)

v 1 -periodic 2-exponents of SU(2 e ) and SU(2 e + 1) Joural of Pure ad Appled Algebra 216 (2012) 1268 1272 Cotets lsts avalable at ScVerse SceceDrect Joural of Pure ad Appled Algebra joural homepage: www.elsever.com/locate/jpaa v 1 -perodc 2-expoets of SU(2

More information

A NEW LOG-NORMAL DISTRIBUTION

A NEW LOG-NORMAL DISTRIBUTION Joural of Statstcs: Advaces Theory ad Applcatos Volume 6, Number, 06, Pages 93-04 Avalable at http://scetfcadvaces.co. DOI: http://dx.do.org/0.864/jsata_700705 A NEW LOG-NORMAL DISTRIBUTION Departmet of

More information

Research Article Multidimensional Hilbert-Type Inequalities with a Homogeneous Kernel

Research Article Multidimensional Hilbert-Type Inequalities with a Homogeneous Kernel Hdaw Publshg Corporato Joural of Iequaltes ad Applcatos Volume 29, Artcle ID 3958, 2 pages do:.55/29/3958 Research Artcle Multdmesoal Hlbert-Type Iequaltes wth a Homogeeous Kerel Predrag Vuovć Faculty

More information

Hypersurfaces with Constant Scalar Curvature in a Hyperbolic Space Form

Hypersurfaces with Constant Scalar Curvature in a Hyperbolic Space Form Hypersurfaces wth Costat Scalar Curvature a Hyperbolc Space Form Lu Xm ad Su Wehog Abstract Let M be a complete hypersurface wth costat ormalzed scalar curvature R a hyperbolc space form H +1. We prove

More information

On L- Fuzzy Sets. T. Rama Rao, Ch. Prabhakara Rao, Dawit Solomon And Derso Abeje.

On L- Fuzzy Sets. T. Rama Rao, Ch. Prabhakara Rao, Dawit Solomon And Derso Abeje. Iteratoal Joural of Fuzzy Mathematcs ad Systems. ISSN 2248-9940 Volume 3, Number 5 (2013), pp. 375-379 Research Ida Publcatos http://www.rpublcato.com O L- Fuzzy Sets T. Rama Rao, Ch. Prabhakara Rao, Dawt

More information

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines It J Cotemp Math Sceces, Vol 5, 2010, o 19, 921-929 Solvg Costraed Flow-Shop Schedulg Problems wth Three Maches P Pada ad P Rajedra Departmet of Mathematcs, School of Advaced Sceces, VIT Uversty, Vellore-632

More information

ECON 5360 Class Notes GMM

ECON 5360 Class Notes GMM ECON 560 Class Notes GMM Geeralzed Method of Momets (GMM) I beg by outlg the classcal method of momets techque (Fsher, 95) ad the proceed to geeralzed method of momets (Hase, 98).. radtoal Method of Momets

More information

A tighter lower bound on the circuit size of the hardest Boolean functions

A tighter lower bound on the circuit size of the hardest Boolean functions Electroc Colloquum o Computatoal Complexty, Report No. 86 2011) A tghter lower boud o the crcut sze of the hardest Boolea fuctos Masak Yamamoto Abstract I [IPL2005], Fradse ad Mlterse mproved bouds o the

More information

Large and Moderate Deviation Principles for Kernel Distribution Estimator

Large and Moderate Deviation Principles for Kernel Distribution Estimator Iteratoal Mathematcal Forum, Vol. 9, 2014, o. 18, 871-890 HIKARI Ltd, www.m-hkar.com http://dx.do.org/10.12988/mf.2014.4488 Large ad Moderate Devato Prcples for Kerel Dstrbuto Estmator Yousr Slaou Uversté

More information

A Study on Generalized Generalized Quasi hyperbolic Kac Moody algebra QHGGH of rank 10

A Study on Generalized Generalized Quasi hyperbolic Kac Moody algebra QHGGH of rank 10 Global Joural of Mathematcal Sceces: Theory ad Practcal. ISSN 974-3 Volume 9, Number 3 (7), pp. 43-4 Iteratoal Research Publcato House http://www.rphouse.com A Study o Geeralzed Geeralzed Quas (9) hyperbolc

More information

ON THE STRUCTURE OF THE SPREADING MODELS OF A BANACH SPACE

ON THE STRUCTURE OF THE SPREADING MODELS OF A BANACH SPACE ON THE STRUCTURE OF THE SPREADING MODELS OF A BANACH SPACE G. ANDROULAKIS, E. ODELL, TH. SCHLUMPRECHT, N. TOMCZAK-JAEGERMANN Abstract We study some questos cocerg the structure of the set of spreadg models

More information

On the Approximation Properties of Bernstein Polynomials via Probabilistic Tools

On the Approximation Properties of Bernstein Polynomials via Probabilistic Tools Boletí de la Asocacó Matemátca Veezolaa, Vol. X, No. 1 003 5 O the Approxmato Propertes of Berste Polyomals va Probablstc Tools Heryk Gzyl ad José Lus Palacos Abstract We study two loosely related problems

More information

Solving Interval and Fuzzy Multi Objective. Linear Programming Problem. by Necessarily Efficiency Points

Solving Interval and Fuzzy Multi Objective. Linear Programming Problem. by Necessarily Efficiency Points Iteratoal Mathematcal Forum, 3, 2008, o. 3, 99-06 Solvg Iterval ad Fuzzy Mult Obectve ear Programmg Problem by Necessarly Effcecy Pots Hassa Mshmast Neh ad Marzeh Aleghad Mathematcs Departmet, Faculty

More information

On Submanifolds of an Almost r-paracontact Riemannian Manifold Endowed with a Quarter Symmetric Metric Connection

On Submanifolds of an Almost r-paracontact Riemannian Manifold Endowed with a Quarter Symmetric Metric Connection Theoretcal Mathematcs & Applcatos vol. 4 o. 4 04-7 ISS: 79-9687 prt 79-9709 ole Scepress Ltd 04 O Submafolds of a Almost r-paracotact emaa Mafold Edowed wth a Quarter Symmetrc Metrc Coecto Mob Ahmad Abdullah.

More information

Unimodality Tests for Global Optimization of Single Variable Functions Using Statistical Methods

Unimodality Tests for Global Optimization of Single Variable Functions Using Statistical Methods Malaysa Umodalty Joural Tests of Mathematcal for Global Optmzato Sceces (): of 05 Sgle - 5 Varable (007) Fuctos Usg Statstcal Methods Umodalty Tests for Global Optmzato of Sgle Varable Fuctos Usg Statstcal

More information

E be a set of parameters. A pair FE, is called a soft. A and GB, over X is the soft set HC,, and GB, over X is the soft set HC,, where.

E be a set of parameters. A pair FE, is called a soft. A and GB, over X is the soft set HC,, and GB, over X is the soft set HC,, where. The Exteso of Sgular Homology o the Category of Soft Topologcal Spaces Sad Bayramov Leoard Mdzarshvl Cgdem Guduz (Aras) Departmet of Mathematcs Kafkas Uversty Kars 3600-Turkey Departmet of Mathematcs Georga

More information

UNIT 2 SOLUTION OF ALGEBRAIC AND TRANSCENDENTAL EQUATIONS

UNIT 2 SOLUTION OF ALGEBRAIC AND TRANSCENDENTAL EQUATIONS Numercal Computg -I UNIT SOLUTION OF ALGEBRAIC AND TRANSCENDENTAL EQUATIONS Structure Page Nos..0 Itroducto 6. Objectves 7. Ital Approxmato to a Root 7. Bsecto Method 8.. Error Aalyss 9.4 Regula Fals Method

More information

ONE GENERALIZED INEQUALITY FOR CONVEX FUNCTIONS ON THE TRIANGLE

ONE GENERALIZED INEQUALITY FOR CONVEX FUNCTIONS ON THE TRIANGLE Joural of Pure ad Appled Mathematcs: Advaces ad Applcatos Volume 4 Number 205 Pages 77-87 Avalable at http://scetfcadvaces.co. DOI: http://.do.org/0.8642/jpamaa_7002534 ONE GENERALIZED INEQUALITY FOR CONVEX

More information

Feature Selection: Part 2. 1 Greedy Algorithms (continued from the last lecture)

Feature Selection: Part 2. 1 Greedy Algorithms (continued from the last lecture) CSE 546: Mache Learg Lecture 6 Feature Selecto: Part 2 Istructor: Sham Kakade Greedy Algorthms (cotued from the last lecture) There are varety of greedy algorthms ad umerous amg covetos for these algorthms.

More information

The Brunn Minkowski Inequality, Minkowski s First Inequality, and Their Duals 1

The Brunn Minkowski Inequality, Minkowski s First Inequality, and Their Duals 1 Joural of Mathematcal Aalyss ad Applcatos 245, 502 52 2000 do:0.006 jmaa.2000.6774, avalable ole at http: www.dealbrary.com o The Bru Mkowsk Iequalty, Mkowsk s Frst Iequalty, ad Ther Duals R. J. Garder

More information

Exercises for Square-Congruence Modulo n ver 11

Exercises for Square-Congruence Modulo n ver 11 Exercses for Square-Cogruece Modulo ver Let ad ab,.. Mark True or False. a. 3S 30 b. 3S 90 c. 3S 3 d. 3S 4 e. 4S f. 5S g. 0S 55 h. 8S 57. 9S 58 j. S 76 k. 6S 304 l. 47S 5347. Fd the equvalece classes duced

More information

1. A real number x is represented approximately by , and we are told that the relative error is 0.1 %. What is x? Note: There are two answers.

1. A real number x is represented approximately by , and we are told that the relative error is 0.1 %. What is x? Note: There are two answers. PROBLEMS A real umber s represeted appromately by 63, ad we are told that the relatve error s % What s? Note: There are two aswers Ht : Recall that % relatve error s What s the relatve error volved roudg

More information

Introduction to Probability

Introduction to Probability Itroducto to Probablty Nader H Bshouty Departmet of Computer Scece Techo 32000 Israel e-mal: bshouty@cstechoacl 1 Combatorcs 11 Smple Rules I Combatorcs The rule of sum says that the umber of ways to choose

More information

Ideal multigrades with trigonometric coefficients

Ideal multigrades with trigonometric coefficients Ideal multgrades wth trgoometrc coeffcets Zarathustra Brady December 13, 010 1 The problem A (, k) multgrade s defed as a par of dstct sets of tegers such that (a 1,..., a ; b 1,..., b ) a j = =1 for all

More information

Special Instructions / Useful Data

Special Instructions / Useful Data JAM 6 Set of all real umbers P A..d. B, p Posso Specal Istructos / Useful Data x,, :,,, x x Probablty of a evet A Idepedetly ad detcally dstrbuted Bomal dstrbuto wth parameters ad p Posso dstrbuto wth

More information

Numerical Simulations of the Complex Modied Korteweg-de Vries Equation. Thiab R. Taha. The University of Georgia. Abstract

Numerical Simulations of the Complex Modied Korteweg-de Vries Equation. Thiab R. Taha. The University of Georgia. Abstract Numercal Smulatos of the Complex Moded Korteweg-de Vres Equato Thab R. Taha Computer Scece Departmet The Uversty of Georga Athes, GA 002 USA Tel 0-542-2911 e-mal thab@cs.uga.edu Abstract I ths paper mplemetatos

More information

The Occupancy and Coupon Collector problems

The Occupancy and Coupon Collector problems Chapter 4 The Occupacy ad Coupo Collector problems By Sarel Har-Peled, Jauary 9, 08 4 Prelmares [ Defto 4 Varace ad Stadard Devato For a radom varable X, let V E [ X [ µ X deote the varace of X, where

More information

Dimensionality Reduction and Learning

Dimensionality Reduction and Learning CMSC 35900 (Sprg 009) Large Scale Learg Lecture: 3 Dmesoalty Reducto ad Learg Istructors: Sham Kakade ad Greg Shakharovch L Supervsed Methods ad Dmesoalty Reducto The theme of these two lectures s that

More information

A conic cutting surface method for linear-quadraticsemidefinite

A conic cutting surface method for linear-quadraticsemidefinite A coc cuttg surface method for lear-quadratcsemdefte programmg Mohammad R. Osoorouch Calfora State Uversty Sa Marcos Sa Marcos, CA Jot wor wth Joh E. Mtchell RPI July 3, 2008 Outle: Secod-order coe: defto

More information

Analysis of Lagrange Interpolation Formula

Analysis of Lagrange Interpolation Formula P IJISET - Iteratoal Joural of Iovatve Scece, Egeerg & Techology, Vol. Issue, December 4. www.jset.com ISS 348 7968 Aalyss of Lagrage Iterpolato Formula Vjay Dahya PDepartmet of MathematcsMaharaja Surajmal

More information

Cubic Nonpolynomial Spline Approach to the Solution of a Second Order Two-Point Boundary Value Problem

Cubic Nonpolynomial Spline Approach to the Solution of a Second Order Two-Point Boundary Value Problem Joural of Amerca Scece ;6( Cubc Nopolyomal Sple Approach to the Soluto of a Secod Order Two-Pot Boudary Value Problem W.K. Zahra, F.A. Abd El-Salam, A.A. El-Sabbagh ad Z.A. ZAk * Departmet of Egeerg athematcs

More information

Extend the Borel-Cantelli Lemma to Sequences of. Non-Independent Random Variables

Extend the Borel-Cantelli Lemma to Sequences of. Non-Independent Random Variables ppled Mathematcal Sceces, Vol 4, 00, o 3, 637-64 xted the Borel-Catell Lemma to Sequeces of No-Idepedet Radom Varables olah Der Departmet of Statstc, Scece ad Research Campus zad Uversty of Tehra-Ira der53@gmalcom

More information

CIS 800/002 The Algorithmic Foundations of Data Privacy October 13, Lecture 9. Database Update Algorithms: Multiplicative Weights

CIS 800/002 The Algorithmic Foundations of Data Privacy October 13, Lecture 9. Database Update Algorithms: Multiplicative Weights CIS 800/002 The Algorthmc Foudatos of Data Prvacy October 13, 2011 Lecturer: Aaro Roth Lecture 9 Scrbe: Aaro Roth Database Update Algorthms: Multplcatve Weghts We ll recall aga) some deftos from last tme:

More information

PPCP: The Proofs. 1 Notations and Assumptions. Maxim Likhachev Computer and Information Science University of Pennsylvania Philadelphia, PA 19104

PPCP: The Proofs. 1 Notations and Assumptions. Maxim Likhachev Computer and Information Science University of Pennsylvania Philadelphia, PA 19104 PPCP: The Proofs Maxm Lkhachev Computer ad Iformato Scece Uversty of Pesylvaa Phladelpha, PA 19104 maxml@seas.upe.edu Athoy Stetz The Robotcs Isttute Carege Mello Uversty Pttsburgh, PA 15213 axs@rec.r.cmu.edu

More information

Arithmetic Mean and Geometric Mean

Arithmetic Mean and Geometric Mean Acta Mathematca Ntresa Vol, No, p 43 48 ISSN 453-6083 Arthmetc Mea ad Geometrc Mea Mare Varga a * Peter Mchalča b a Departmet of Mathematcs, Faculty of Natural Sceces, Costate the Phlosopher Uversty Ntra,

More information

Packing of graphs with small product of sizes

Packing of graphs with small product of sizes Joural of Combatoral Theory, Seres B 98 (008) 4 45 www.elsever.com/locate/jctb Note Packg of graphs wth small product of szes Alexadr V. Kostochka a,b,,gexyu c, a Departmet of Mathematcs, Uversty of Illos,

More information

2006 Jamie Trahan, Autar Kaw, Kevin Martin University of South Florida United States of America

2006 Jamie Trahan, Autar Kaw, Kevin Martin University of South Florida United States of America SOLUTION OF SYSTEMS OF SIMULTANEOUS LINEAR EQUATIONS Gauss-Sedel Method 006 Jame Traha, Autar Kaw, Kev Mart Uversty of South Florda Uted States of Amerca kaw@eg.usf.edu Itroducto Ths worksheet demostrates

More information

AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET

AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET Abstract. The Permaet versus Determat problem s the followg: Gve a matrx X of determates over a feld of characterstc dfferet from

More information

ON THE LOGARITHMIC INTEGRAL

ON THE LOGARITHMIC INTEGRAL Hacettepe Joural of Mathematcs ad Statstcs Volume 39(3) (21), 393 41 ON THE LOGARITHMIC INTEGRAL Bra Fsher ad Bljaa Jolevska-Tueska Receved 29:9 :29 : Accepted 2 :3 :21 Abstract The logarthmc tegral l(x)

More information