CONSTRUCTING GENERALIZED MEAN FUNCTIONS USING CONVEX FUNCTIONS WITH REGULARITY CONDITIONS

Size: px
Start display at page:

Download "CONSTRUCTING GENERALIZED MEAN FUNCTIONS USING CONVEX FUNCTIONS WITH REGULARITY CONDITIONS"

Transcription

1 CONSTRUCTING GENERALIZED MEAN FUNCTIONS USING CONVEX FUNCTIONS WITH REGULARITY CONDITIONS YUN-BIN ZHAO, SHU-CHERNG FANG, AND DUAN LI Abstract. The geeralzed mea fucto has bee wdely used covex aalyss ad mathematcal programmg. Ths paper studes a further geeralzato of such a fucto. A ecessary ad suffcet codto s obtaed for the covexty of a geeralzed fucto. Addtoal suffcet codtos that ca be easly checked are derved for the purpose of detfyg some classes of fuctos whch guaratee the covexty of the geeralzed fuctos. We show that some ew classes of covex fuctos wth certa regularty such as S -regularty ca be used as buldg blocks to costruct such geeralzed fuctos. Key words. Covexty, mathematcal programmg, geeralzed mea fucto, self-cocordat fuctos, S -regular fuctos. AMS subject classfcatos. 90C30, 90C25, 52A4, 49J52. Itroducto. I ths paper, we deote the -dmesoal Eucldea space by R, ts oegatve orthat by R+, ad postve orthat by R++. I 934, Hardy, Lttlewood ad Pólya [3] cosdered the followg fucto uder the ame of geeralzed mea:. Υ w x = φ w φx where φ s a real, strctly creasg, covex fucto defed o a subset of R ad w = w, w 2,..., w T s a gve vector R+. Assumg that φ > 0, φ > 0 ad > 0, they showed a equvalet codto for the covexty of Υ w. Whe φ s three tmes dfferetable, Be-Tal ad Teboulle [2] establshed aother equvalet codto for Υ w beg covex see ext secto for detals. The geeralzed mea fucto. has may applcatos optmzato. Be- Tal ad Teboule [2] demostrated a terestg applcato of. a cotuous form o pealty fuctos ad dualty formulato of stochastc olear programmg problems. However, the most wdely used geeralzed meas are the logorthmc-expoetoal ad p-orm fuctos: /p f w x = log w e x, p w x = w x p for x = x,..., x T R. They correspod to the specal cases of Υ w wth φt = e t ad φt = t p, respectvely. Needless to say that the log-exp fucto has bee wdely used covex aalyss ad mathematcal programmg. For example, a geometrc program see Duff et Isttute of Appled Mathematcs, AMSS, Chese Academy of Sceces, Bejg 00080, Cha Emal: ybzhao@amss.ac.c. Ths author s work was partally supported by the Natoal Natural Scece Foudato of Cha uder Grat No ad Grat No Correspodg author. Idustral Egeerg ad Operatos Research, North Carola State Uversty, Ralegh, NC , USA Emal: fag@eos.csu.edu. Ths author s work was partally supported by the US Army Research Offce Grat No. W9NF-04-D Departmet of Systems Egeerg ad Egeerg Maagemet, Chese Uversty of Hog Kog, Shat, NT, Hog Kog Emal: dl@se.cuhk.edu.hk. Ths author s work was partally supported by Grat CUHK480/03E, Research Grat Coucl, Hog Kog.

2 al. [8] ad Boyd ad Vadeberghe [6] ca be coverted to a covex programmg problem by usg the log-exp fucto so that the teror-pot algorthms ca be developed to solve geometrc programs wth great effcecy Kortaek et al. [4]. Aother example s cocered wth the odfferetable mmax problem m max g y, y D where g, =,...,, are real fuctos defed o a covex set D R m. Sce the recesso fucto of the log-exp fucto s the max-fucto see Rockafellar [20],.e., max x = lm ε 0+ εf x ε where f = f w ad w =,,...,, the above odfferetal optmzato problem ca be approxmated by solvg the followg optmzato problem m ε log e g y ε. y D Ths objectve fucto s dfferetable ad covex, f every g y s. Other applcatos of the log-exp fucto optmzato ca be foud Be-Tal [], Be-Tal ad Teboulle [3], Zag [25], Bersekas [4], Polyak [9], Fag [9, 0], L ad Fag [5], Peg ad L [7], Brbl et al. [5], Su ad L [22, 23, 24], etc. It s worth metog that the cojugate fucto of the log-exp fucto happes to be the well-kow Shao s etropy fucto [2] whch plays a vtal role so may felds ragg from the mage ehacemet to ecoomcs ad from statstcal mechacs to uclear physcs see, Buck ad Macaulay [7] ad Fag et al. []. We cosder ths paper a further geeralzato of. the followg form:.2 Γ w x = Ψ w φ x where φ : Ω R, =,...,, are covex, twce dfferetable but ot ecessarly beg strctly creasg fuctos defed o a ope covex set Ω R, Ψ : Ω R s covex, twce dfferetable ad strctly creasg, ad w R+ s a gve vector. Clearly, Υ w s a specal case of Γ w wth φ = φ 2 =... = φ = Ψ = φ. For coveece, ths paper, we stll call Γ w gve by.2 a geeralzed mea fucto, ad we call φ the er fucto ad Ψ the outer fucto of Γ w. To assure the well-defedess of Γ w, we aturally requre that Coe[φ Ω] ΨΩ, where Coe[φ Ω] deotes the coe geerated by the set φ Ω. As the case of Υ w, we would lke to derve certa suffcet ad ecessary codtos for the fucto Γ w to be covex. Moreover, we hope to fd a systematc way to explctly costruct some classes of covex Γ w. It s terestg to pot out that Γ w s by o meas a ew research subject. I fact, t was essetally studed by W. Fechel hs lecture otes of Covex Coes, Sets ad Fuctos 953 [2]. Based o the propertes of level sets ad characterstc roots of Hessa matrces of fuctos volved, Fechel derved some suffcet ad ecessary codtos for the covexty of the geeralzed mea fucto Γ w. The codtos he derved, however, are rather complcated, ad there s o smple test to decde what kd of fuctos may admt these complcated propertes. Ulke Fechel s approach, our aalyss ths paper depeds oly o the fucto value, ts frst dervatve, ad secod dervatve to provde a suffcet ad ecessary codto for Γ w beg covex. The ecessary ad suffcet codto we derve ths paper 2

3 ca be vewed as a geeralzato of that [3] cocerg the fucto.. We ca also use related suffcet codtos to explctly costruct cocrete examples of covex Γ w. Moreover, we show how the so-called S -regular fuctos to be defed ths paper ca be used to costruct covex geeralzed mea fuctos. The rest of the paper s orgazed as follows. I Secto 2, we vestgate the codtos that assure the covexty of the geeralzed mea fucto Γ w. I Secto 3, we detfy some classes of fuctos that satsfy the codtos derved Secto 2, ad llustrate how the geeralzed mea fucto Γ w ca be explctly costructed. Coclusos are gve the last secto. 2. Necessary ad Suffcet Codtos for the Covexty of Γ w. Let us start wth a smple lemma proof omtted that shows the verse of a creasg covex fucto s cocave ad creasg. Lemma 2.. Let Ω be a ope covex subset of R ad Ψ : Ω R be a real fucto defed o Ω. The Ψ s strctly covex ad strctly creasg f ad oly f ts verse Ψ : R Ω s strctly cocave ad strctly creasg. Notce that f w = 0, for some, the the term w φ x ca be removed from the expresso of Γ w x, ad t suffces to cosder Γ w defed o R. Thus, wthout loss of geeralty, we may assume that the vector w R++ throughout the rest of the paper. To study the covexty of Γ w, whe assumg that φ, =,...,, ad Ψ are twce dfferetable, we eed to check the propertes of ts Hessa matrx. Let Sce xw x Moreover, = w φ x, we have x w = w φ x. Γ w x = Ψ x w w φ x. 2 Γ w x 2 = Ψ x w w φ x 2 + Ψ x w w x, 2 Γ w x x j = Ψ x w w w j φ x φ jx j for j. Cosequetly, the Hessa matrces of Γ w becomes w x Γ w x 2 = Ψ x w 0 w 2 2x w x w φ x + Ψ x w w 2 φ 2x [w φ x, w 2 φ 2x 2,..., w φ x ]. w φ x Note that whe φ, =,...,, are all covex ad Ψ s covex ad creasg, by Lemma 2., we see that the frst term o the rght-had sde of 2. s a postve 3

4 semdefte matrx multpled by a postve coeffcet Ψ x w, whle the secod s a rak oe matrx multpled by a egatve coeffcet Ψ x w. Some codtos for covexty of the fucto Υ w x has already bee studed [3] ad [2]. We summarze ther results here. Theorem 2.2. [3] Uder the codtos of φ > 0, φ > 0 ad > 0, the fucto Υ w x defed by. s covex f ad oly f the followg codto holds: w [φ x ] 2 x [φ y] 2 y for y = Υ w x. Be-Tal ad Teboulle [2] also provded a dfferet suffcet ad ecessary codto, uder certa assumptos, for the covexty of the fucto Υ w x. Theorem 2.3. [2] Let φt C 3. Υ w x s covex f ad oly f /ρt s covex, where ρt = /φ. It s possble to exted the aalyss [2] to derve suffcet codtos for more geeral stuatos where the covexty of Γ w x defed by.2 s cosdered. Actually, the followg result ca be proved alog the le of the proof of Lemma ad Theorem [2]. Theorem 2.4. [2] Let Ψt C 3 ad φ t C 3 be strctly creasg ad ρt = Ψ t/ψ t. If /ρt s covex ad Ψ φ t s covex, for =,...,, the Γ w x gve by.2 s covex. Note that f φ s suffcetly smooth, /ρt s covex, where ρt = t/φ t, f ad oly f ts secod dervatve s oegatve,.e., = φ 3 2φ 2 + φ 2 ρ 2 0. Thus, to check the covexty of /ρt, t s usually eeded to check the above equalty volvg the thrd ad forth dervatve of the fucto φ. Note that Theorem 2.2, however, does ot requre the thrd or fourth dfferetablty of the fucto φ. I what follows, we geeralze the results Theorem 2.2 for the fucto Γ w x. Although the basc dea of our proof s closely related to that of [3], the proof s ot straghtforward. For completeess, we gve a detaled proof for the result. Theorem 2.5. Let Ω R be ope ad covex, Ψ : Ω R be covex, twce dfferetable ad strctly creasg, φ : Ω R, =,...,, be strctly covex ad twce dfferetable, ad w R++ be a gve vector. The the geeralzed mea fucto Γ w x = Ψ w φ x {}}{ s covex o Ω := Ω Ω f ad oly f Ψ [φ y w x ] x [Ψ y] 2 for x Ω ad y = Γ w x. Moreover, Γ w x s strctly covex f ad oly f the equalty 2.2 holds strctly. Proof. Let y = Γ w x = Ψ x w. The x w = Ψy ad 2.3 Ψ x w Ψ y =. 4

5 Dfferetatg both sdes wth respect to y ad use the above relatos, we have Therefore, 0 = Ψ x w [Ψ y] 2 + Ψ x w Ψ y = Ψ x w [Ψ y] 2 + Ψ y Ψ y. 2.4 Ψ x w = Ψ y [Ψ y] 3. Combg 2.3 ad 2.4 yelds 2.5Ψ x w + Ψ x w [φ w x ] 2 [Ψ y] 2 x = w [φ x] 2 Ψ y x. [Ψ y] 3 Frst we prove that Γ w x s covex, f 2.2 holds. It suffces to show that the Hessa matrx of Γ w x s postve sem-defte. For ay d R ad x Ω, the Cauchy-Schwartz equalty mples that 2 w φ x d = [ ] w x d w x d 2 w x φ x w [φ x ] 2 x By Lemma 2., we kow Ψ s cocave ad hece Ψ x w 0 for all x w. Combg ths fact wth the above equalty, we see that, for ay d R, 2. d T 2 Γ w x 2 d 2 = Ψ x w w x d 2 + Ψ x w w φ x d Ψ x w = = 0. w x d 2 w x d 2 w x d 2 [ + Ψ x w Ψ x w + Ψ x w [Ψ y] 2 w [φ x]2 x [Ψ y] 3 w x d 2 Ψ y w [φ x ] 2 x ] [φ w x ] 2 x The last equalty follows from 2.5 ad the last equalty follows from the fact that the frst quatty o the rght-had sde,.e., w x d 2, s oegatve, ad the secod quatty s also oegatve due to our assumpto. Cosequetly, we have prove that the Hessa matrx 2 Γ w x s postve sem-defte, as desred. 2 5

6 Coversely, we would lke to show that equalty 2.2 holds, f Γ w x s covex. For ay vector 0 d R, kowg 2.3, 2.4 ad the covexty of Γ w x, we have 0 d T 2 Γ w x 2 d = 2 Ψ x w w x d 2 + Ψ x w w φ x d = 2 Ψ w x d 2 Ψ y y Ψ y 3 w φ x d [ ] = w x d 2 Ψ y Ψ y [ w φ x d ] Ψ y 3 w x d 2. Notce that the above equalty holds for ay vector d R. I partcular, let The, we have d = w φ x d =, φ x x k= w [φ k x k] 2 k k x k w x d 2 = As a result, the equalty 2.6 reduces to [ 0 w [φ x]2 Ψ y Ψ y Ψ y 3 x, =,...,. w. [φ x]2 x ] [φ w x ] 2 x We see that equalty 2.2 deed holds. The result about strct covexty ca be easly checked out. Theorem 2.5 geeralzes the result of Theorem 2.2 cocerg Υ w x to the more geeral fucto Γ w x, whle Theorem 2.4 geeralzes the suffcet codto of Theorem 2.3 cocerg Υ w x to the fucto Γ w x, uder dfferet assumptos. Except for some very smple cases, such as e t or x p, these results, however, do ot gve us a cocrete class of fuctos whch ca be used to costruct specfc geeralzed mea fuctos. The purpose of the remader of ths paper s to provde a systematc way to detfy the desred class of fuctos. Our aalyss here the paper s based oly o the result of Theorem 2.5. We beleve that there should also exst some parallel results based o Theorem 2.4. To ths ed, two related suffcecy results of Theorem 2.5 are derved below for ther coveet usage costructg covex Γ w see ext secto. Theorem 2.6. Let Ω be a ope covex subset of R, Ψ : Ω R be strctly creasg, twce dfferetable ad covex, φ : Ω R, =,...,, be strctly covex ad twce dfferetable, ad w R++ be a gve vector. Assume that there exsts a scalar α R such that 2.7 αψtψ t [Ψ t] 2 for t Ω. The the fucto Γ w s covex o Ω f 2.8 w [φ x ] 2 x αψy for x Ω, 6

7 where y = Γ w x. Proof. Multplyg both sdes of 2.8 by Ψ y ad applyg 2.7, we see that codto 2.2 holds. The result follows from Theorem 2.5 mmedately. Theorem 2.7. Let Ω be a ope covex subset of R, Ψ : Ω R be strctly creasg, twce dfferetable ad covex, φ : Ω R, =,...,, be strctly covex ad twce dfferetable, ad w R++ be a gve vector. Assume that there exst 0 α R, =,...,, holdg the same sg such that 2.9 α φ t t [φ t] 2 for t Ω, ad there exsts a α R such that the equalty 2.7 holds. The the fucto Γ w s covex f 2.0 or 2. α max α α m α whe α > 0 for all, whe α < 0 for all. Proof. Takg y = Γ w x, we see two cases. Case : α > 0 for =,...,. I ths case, 2.9 mples that φ t 0 for t Ω ad 2.0 mples that [φ w t]2 x w α φ x max α w φ x αψy. Case 2: α < 0 for =,...,. I ths case, 2.9 mples that φ t 0 for t Ω ad 2. mples that [φ w t]2 x w α φ x m α w φ x αψy. Both cases yeld 2.8 ad the desred result follows from Theorem 2.2. A specal case of φ t = φ 2 t =... = Ψt mmedately leads to the ext result. Corollary 2.8. Let Ω be a ope covex set R, φ : Ω R be a covex, twce dfferetable ad strctly creasg fucto, ad w R++ be a gve vector. If there exsts a α 0 such that 2.2 [φ t] 2 = αφt t for t Ω, the the fucto Υ w x = φ w φx s covex o Ω. Ths result ca also follow drectly from the aforemetoed Theorem 2.3 due to Be-Tal ad Teboulle [2]. I fact, t s easy to verfy that the relato 2.2 mples that the secod dervatve of φ / s equal to zero, ad thus by Theorem 2.3 the fucto Υ w x s covex. Remark 2.. The fuctos satsfyg a dfferetal equalty such as 2.7 are related to the so-called self-cocordat barrer fucto troduced by Nesterov ad Nemrovsky [6]. Recall that a C 3 fucto ξ : 0, R s sad to be self-cocordat f ξ s covex ad there exsts a costat µ > 0 such that 2.3 ξ t µ ξ t 3 2 for t 0,. 7

8 Moreover, the self-cocordat fucto ξ s called a self-cocordat barrer fucto f there exsts a costat µ 2 > 0 such that 2.4 ξ t µ 2 [f t] 2 for t 0,. Combg 2.3 ad 2.4 yelds ξ tξ t µ[ξ t] 2. Ths dcates that the frst-order dervatve fucto of a self-cocordat barrer fucto,.e., gt := ξ t, satsfes the equalty 2.7. A self-cocordat fucto ξ tself may also satsfy a equalty lke 2.7 or 2.9. Remark 2.2. The fuctos satsfyg a dfferetal equalty such as 2.7 also appear covexty theory. doma Ω, we cosder the covexty of the fucto ht := φt Gve a twce dfferetable fucto φt > 0 o ts o Ω. Notce that h t = 2[φ t] 2 φt t [φt] 3 for t Ω. Hece the fucto ht = φt s covex f ad oly f the equalty φtφ t 2[φ t] 2 holds o Ω. Moreover, f φt t [φ t] 2, the covex fucto ht satsfes a reverse equalty,.e., hth t [h t] 2 o Ω. From ths observato, a related questo arses. Gve a fucto φt > 0 o Ω ad a costat r > 0, whe wll the fucto ht := φt become covex ad satsfy r a equalty such as 2.9? A straghtforward aalyss leads to the ext result. Theorem 2.9. Let Ω be a covex subset of R ad φ : Ω 0, be a fucto. If φt t [φ t] 2 for t Ω, the, for ay r > 0, the fucto ht := φt s r covex ad hth t [h t] 2 for t Ω. Coversely, f there exsts a r > 0 such that ht := φt s covex ad hth t [h t] 2 for t Ω, the φt t [φ t] 2 r for t Ω. Let Ω be a covex subset of R, τ > 0, ad φ : Ω τ, be a fucto. If φt t [φ t] 2 for t Ω, the, for ay scalar r > 0 ad T > 0, the fucto h T t := T + φt s covex ad αh r T th T t [h T t]2 for t Ω, where α = T τ r +. Proof. For case, t s suffcet to see that ad h t = r2 φ t 2 + r[φ t 2 φt t] φt r+2, hth t [h t] 2 = r[φ t 2 φt t] φt 2r+. For case, t s easy to verfy that h T t = h t ad T φt r h T th T t [h T t] 2 = r[φ t 2 φt t]. + φt 2r+ The the desred result follows. The above results dcate that f we have a fucto φ satsfyg the equalty 2.7 wth α =, the we may costruct a fucto h from φ such that h satsfes the 8

9 coverse dfferetable equalty αhth t [h t] 2 for some costat α. Moreover, f we take a T-traslato of the value of the fucto h, the the resultg fucto satsfes the coverse dfferetable equalty wth a α that ca be reduced to be smaller tha ay threshold gve 0, provded a sutable choce of T > 0. Ths fact wll be used ear the ed of Secto Costructg covex geeralzed mea fuctos Γ w. I ths secto, we develop procedures to detfy some classes of fuctos that satsfy equalty 2.7 ad/or equalty 2.9 so that we have buldg blocks for costructg cocrete covex fucto Γ w x. Frst, we gve a result that detfes fuctos satsfyg the equato 2.2. Obvously, ths class of fuctos satsfes both equaltes 2.7 ad 2.9. Theorem 3.. Let Ω be a ope set R ad φ : Ω R be a covex, twce dfferetable ad strctly creasg fucto satsfyg equato 2.2 wth a costat α 0. The, whe α =, φ s the form of φt = γe t β for some γ > 0 ad β > 0. whe 0 < α wth v α := sup t Ω α t beg fte, φ s the form of α α φt = γ α t + β α for some γ > 0 ad β v. whe α < 0 wth u α := sup t Ω α t beg fte, φ s the form of φt = γ β α α α t α for some γ > 0 ad β u. Note that results ad were poted out [2] ad [3] ad result ca be easly derved. The above result leads to the followg cosequece related to Υ w. Corollary 3.2. The followg fuctos ca be used to explctly costruct a covex geeralzed mea fucto Υ w x = φ w φx over Ω : φt = γe t β over Ω = R wth γ > 0 ad β > 0. φt = γ φt = p p t + β over Ω = η, wth p >, γ > 0 ad β η γ over Ω =, η wth p > 0, γ > 0 ad β η β p tp p. v φt = γβ p tp over Ω =, η wth 0 < p <, γ > 0 ad β η p. Aga, results ad were gve [2] ad [3] ad results ad v ca be easly derved. The fuctos lsted Corollary 3.2 actually form a complete bass the sese that the fucto φ case satsfes codto 2.2 wth α = ; the fucto φ case satsfes codto 2.2 wth α = > ; the fucto φ case satsfes codto 2.2 wth α = p p+ 0, ; ad the fucto φ v satsfes codto 2.2 wth α = p p < 0. We ow try to detfy some class of fuctos that satsfy equaltes 2.7 ad/or 2.9. For smplcty, we oly cosder covex, twce dfferetable, strctly creasg fuctos ϑ o Ω = 0,. Let us frst defe the followg four categores of such fuctos: p p U = {ϑ : There exsts α R such that αϑtϑ t [ϑ t] 2 for t Ω}; U 2 = {ϑ : There exsts α R such that αϑtϑ t [ϑ t] 2 for t Ω}; 9 p.

10 U 3 = {ϑ : There exst α α 2 such that α ϑtϑ t [ϑ t] 2 α 2 ϑtϑ t for t Ω}; U 4 = {ϑ : There exsts α R such that αϑtϑ t = [ϑ t] 2 for all t Ω}. It s evdet that U 4 U 3 U 2 U. As poted out Theorem 3., the class U 4 ca be gve explctly. By allowg α α 2, we show that U 3 s much broader tha U 4. I fact, may covex fuctos wth certa regulartes fall to the category U 3. To start, we troduce a ew class of fuctos wth certa regularty propertes. Defto 3.3. A covex, twce dfferetable, strctly creasg fucto δt : 0, R s called a S -regular fucto f δt vashes at t = 0 the sese of lm δ0 = lm δ 0 = lm δ 0 = 0; t 0 + t 0 + t 0 + ad there exst postve costats 0 < β β 2, p ad q such that 3. β [t + p t + q ] δ t β 2 [t + p t + q ], t > 0. Note that codto 3. actually mples the strct covexty of a S -regular fucto o 0,. I partcular, settg β = β 2, codto 3. reduces to a equato 3.2 δ t = t + p t + q. Takg tegrato twce ad otg that lm t 0+ δ0 = lm t 0+ δ 0 = 0, the uque soluto to equato 3.2 s 3.3 p,q t = t + p+ pp + t + q qq p + q t for p ad q >. pq I addto, sce lm q + [ t + q ]/q = lt +, we have 3.4 p, t = t + p+ pp + + lt + p + t for p. p Takg p = 3.4, we have 3.5, t = t Moreover, takg p = ad q = yelds + lt + 2t = 2 t2 t + lt ,2 t = 2 [ t + 2 t + 3t ]. I terms of ths partcular soluto p,q t, codto 3. ca be wrtte as 3.7 β p,qt δ t β 2 p,qt. By tegratg ad otg that lm t 0+ δ 0 = lm t 0+ δ0 = 0, we further have 3.8 β p,qt δ t β 2 p,qt 0

11 ad 3.9 β p,q t δt β 2 p,q t. Therefore, we ca see that the class of S -regular fuctos s qute broad. Later, by usg , we show that S -regular fuctos fall to the category U 3. It s worth metog that for ay p, q > cludg the case of q + the S -regular fucto p,q t s ot self-cocordat. I fact, the fucto p,q t does ot satsfy the equalty 2.3 sce δ t 0 ad δ t p + q as t 0 +. S -regular fuctos are somewhat aalogous to but dfferet from the selfregular fuctos defed [8]. As we have metoed above, S -regular fuctos are ot self-cocordat. The class of self-regular fuctos, however, has a large overlap wth self-cocordat fuctos. I what follows, we dsplay the relato amog the frst ad secod dervatves of S -regular fuctos, whch shows that ay S - regular fucto belogs to the category U 3. It should be metoed that the relatos amog the frst ad secod dervatves for self-regular fucto have bee studed [8]. Theorem 3.4. Let δt : 0, R be S -regular o 0,. The there exst c 2 c > 0 such that 3.0 c δtδ t [δ t] 2 c 2 for all t 0,,.e., the fucto δt U 3. Proof. We show that a S -regular fucto p,q t satsfes the property 3.0. Actually, we have p,q t p,qt [ p,qt] 2 = t+ p+ pp+ t+ q qq t+ p p p+q pq t + t+ q q [t + p t + q ] 2. p+q pq Dvdg the umerator ad deomator of the rght-had sde of the above equato by t + 2p = t + p+ t + p, we have p,qt p,qt [ p,qt] 2 = Therefore, 3. t+ p+ pp+ + t+ p+ t+ p+q qq p+qt t+ p+q pqt+ p p+q p qt+ p+q pqt+ p p,q t p,qt lm t [ p,qt] 2 = p p +. Sce p,qt = t + p t + q, we have lm t 0+ p,qt = p + q. Sce p,qt 0, p,qt 0 ad p,q t 0 as t 0 +, we have p,qt 2 [ p,qt 2 ] lm t 0 + = lm p,qt t 0 + [ p,qt] = lm Hece, we have lm t 0 + p,q t 2 p,qt p,qt = lm t 0 + t p,qt p,qt = 2p + q. p,qt p,qt 2[ p,qt] p,qt p,qt = 6p + q.

12 Usg the above relatos, we further have 3.2 p,q t p,qt p,qt p,qt + p,q t p,qt lm t 0 + [ p,qt] 2 = lm t p,qt p,qt = 2 + lm p,q t t p,qt p,qt lm t 0 + p,qt = 2 3. Notce that p,q t > 0, p,qt > 0 ad p,qt > 0 0,. From 3. ad 3.2, we ca see by cotuty that there exst two costats µ 2 µ > 0 such that µ p,qt p,qt [ p,qt] 2 µ 2 for t 0,. Together wth 3.7 through 3.9, ths mples that a S -regular fucto δt satsfes the followg equalty: 0 < µ β δtδ t [δ t] 2 β 2 µ 2, Therefore, 3.0 holds wth c := µ β ad c 2 := µ 2 β 2. A fact that should be poted out here s that ew fuctos U or U 2 ca be costructed by usg the basc operatos addto, multplcato, dvso ad composto o kow fuctos. The proof of the followg fact s omtted. Lemma 3.5. If φ : 0, 0,, φ U wth α = α ad ϕ : 0, 0,, ϕ U wth α = α 2, the φ + ϕ U wth α = 2 max{α, α 2 }. If φ : 0, 0,, φ U wth α 0, ] ad ϕ : 0, 0,, ϕ U wth α 2 0, ], the the multplcatve fucto φt ϕt U wth α =. Smlarly, f φ U 2 wth α ad ϕ U 2 wth α 2, the φt ϕt U 2 wth α =. If φ : 0, 0,, φ U 2 wth α ad ϕ : 0, 0,, ϕ U wth α 2 0, ], the the fucto φ ϕ U 2 wth α =. Smlarly, f φ U wth α 0, ] ad ϕ U 2 wth α 2, the φ ϕ U wth α =. v Let ϕ : 0, Ω R ad φ : Ω 0, be two covex fuctos. If φ U wth α > 0, the the composte fucto φ ϕt = φϕt U wth the same costat α. The ext result shows that the composte fuctos of e t belog to U 3. Lemma 3.6. Deote the expoetal fucto e t by expt ad the composto of m m expoetal fuctos by θ m t := m {}}{ exp exp expt. The 3.3 m θ mtθ mt [θ mt] 2 θ m tθ mt for t R. Proof. Let α m t := [θ mt] 2 /θ m tθ mt for t R. Sce α t, we ca prove the rght-had sde equalty of 3.3 usg v of Lemma 3.5 ad mathematcal ducto. For the left-had sde equalty, otce that θ mt = θ m tθ m t, θ mt = θ m tθ m t 2 + θ m tθ m t for t R. 2

13 Ths dcates that α m t = + α m tθ m t > + α m t for t R. It s easy to check that α 2 t 2,. The desred result follows by ducto. To costruct examples of the covex fucto Γ w, Theorem 2.7 tells us that t suffces to fd fuctos satsfyg the equaltes 2.7 ad 2.9 ad compare ther α values. The ext result s to estmate the α values, or equvaletly, to estmate the values of c ad c For smplcty, we use the S -regular fuctos wth p = ad q =, 2 to estmate requred c ad c 2. I fact, we have the followg result. Its proof s omtted here. Lemma 3.7. The S -regular fuctos, t ad,2 t gve by 3.5 ad 3.6, respectvely, satsfy codto 3.0 wth c = 2 ad c 2 = 2 3, that s, ,t,t [,t] 2 2, t,t, ,2t,2t [,2t] 2 2,2 t,2t for t 0,. We ow gve the last result o how to costruct some covex fuctos Γ w. Theorem 3.8. Let Ω be a ope covex subset of R. Let φ : Ω 0, be a covex, twce dfferetable, strctly creasg fucto o Ω. If φt t [φ t] 2 for t Ω, the the geeralzed mea fucto Γ w x := φ w φx r s covex o Ω for ay gve w R ++ ad r > 0. Let κ > 0 be a costat ad φ : Ω κ, be a covex, twce dfferetable, strctly creasg fucto satsfyg the equalty φt t [φ t] 2 for t Ω. The, for ay gve w R ++ ad T > 0, r > 0, the fucto {}}{ Γ 2 w x := l l l l w T + φx r s covex o Ω for ay postve teger l T κ r +. I fact, result comes from part of Theorem 2.9 ad Theorem 2.7. Result follows from Lemma 3.6 ad Theorem 2.7, ad part of Theorem 2.9. I fact, t suffces to take the er fucto h T t = T + φt ad outer fucto θ r m t, as m {}}{ l l lt. defed Lemma 3.6, whose verse fucto s gve by The above result partally aswers the followg terestg questo: Gve a covex fucto, how may tmes of log-trasformatos ca be appled whle retag the covexty? Usg Theorems 2.7, 2.9, 3.8 ad Lemma 3.7, we have the followg examples of covex Γ w. 3

14 Example 3.. [ ],j,j x r, l,j x r, l m+ {}}{ l l l m + e rx, x 0,. v l {}}{ l l l [ ] m +, x r, x τ,, l m,τ r +, τ > 0. It follows from Corollary 3.2 that the fucto x p over 0, satsfes 2.2 wth α = p 29 p. Hece, whe < p 2, we have α 2, ad whe < p 7, we have α By Theorems 3.7, both,2t ad, t satsfy codto 2.9 wth α = 2. From Theorem 2.7, we see the fuctos below are examples of covex Γ w. Example 3.2. Let < p 2 ad δ t =,2 t or, t, for t 0, ad =,...,. The Γ w x = w δ x p s covex o 0,. Before closg ths secto, we brefly llustrate a possble applcato of volvg fucto Γ w the regularzato method for solvg a olear programmg problem: m{f 0 x : x C}. For smplcty, we assume that C s a covex set ad f 0 s a covex fucto. Let µ > 0 be a postve parameter. Gve a strctly covex fucto Γ w, we cosder the followg problem: m{f 0 x + µγ w x : x C}. Ths problem becomes a strctly covex programmg problem wth a uque soluto, deoted by xµ, whch comprses of a cotuato trajectory {xµ : µ > 0}. Uder sutable codtos of f 0, Ψ ad φ, ths trajectory becomes bouded. I ths case, by settg µ 0, ay accumulato pot of xµ, as µ 0, s a soluto to the orgal problem. Thus, a path-followg algorthm ca be desged to follow ths trajectory to acheve the soluto of the orgal problem. The performace of such path followg algorthm certaly depeds o the choce of the fucto Γ w wth regularty codtos. 4. Coclusos. I ths paper, we have further exteded the theoretcal foudato for the geeralzed mea fucto. We have establshed a ecessary ad suffcet codto for such a geeralzato to be covex. Moreover, a systematc way to explctly costruct covex Γ w has bee developed. To ths ed, the cocept of S -regular fuctos has bee troduced. It should be oted that ay S -regular fucto s ot self-cocordat [6]. Ackowledgmets. We would lke to thak the two aoymous referees for ther sghtful commets whch help mprove sgfcatly the results ad presetato of the paper. Especally, we are grateful to oe of the referees who brought the refereces [2] ad [3] to our atteto, ad provded Theorem 2.4 ths paper. 4

15 REFERENCES [] A. Be-Tal, The etropc pealty approach to stochastc programmg, Mathematcs of Operatos Research, 0 985, pp [2] A. Be-Tal ad M. Teboulle, Expected utlty, pealty fuctos, ad dualty stochastc olear programmg, Maagemet Scece, , pp [3] A. Be-Tal ad M. Teboulle, A smoothg techque for odfferetable optmzato problems, Optmzato, Lecture Note Mathematcs 405, Sprger Verlag, pp. -, 989. [4] D.P. Bertsekas, Costraed Optmzato ad Lagraga Multpler Methods, Academc Press, New York, 982. [5] S.I. Brbl, S.-C. Fag, J. Frek ad S. Zhag, Recursve approxmate of the hgh dmesoal MAX fucto, Operatos Research Letters, , pp [6] S. Boyd ad L. Vadeberghe, Itroducto to Covex Optmzato wth Egeerg Applcatos, Staford Uversty, Staford, CA, 997. [7] B. Buck ad V.A. Macaulay, Maxmum Etropy Acto: A Collecto of Expostory Essays, Oxford Uversty Press, Lodo UK, 99. [8] R.J. Duff, E.L. Peterso ad C. Zeer, Geometrc Programmg - Theory ad Applcatos, Wley, 967. [9] S.-C. Fag, A ucostraed covex programmg vew of lear programmg, Mathematcal Methods of Operatos Research, , pp [0] S.-C. Fag ad H. S. J. Tsao, O the etropc perturbato ad expoetal pealty methods for lear programmg, Joural of Optmzato Theory ad Applcatos, , pp [] S.-C. Fag, J.R. Rajasekera ad H. Tsao, Etropy Optmzato ad Mathematcal Programmg, Kluwer Academc Publshers, Bosto, MA, 997. [2] W. Fechel, Covex Sets, Coes, ad Fuctos, Lectures Preceto Uversty, Prceto, New Jersey, 953. [3] G. Hardy, J.L. Lttlewood ad G. Pólya, Iequaltes, Cambrdge Uversty Press, 934. [4] K.O. Kortaek, X. Xu ad Y. Ye, A feasble teror-pot algorthm for solvg prmal ad dual geometrc programs, Mathematcal Programmg, , pp [5] X.-S. L ad S.-C. Fag, O the regularzato method for solvg m-max problems wth applcatos, Mathematcal Methods of Operatos Research, , pp [6] Y. Nesterov ad A. Nemrovsky, Iteror-pot Polyomal Methods Covex Programmg, SIAM Studes 3, Phladelpha, PA, 994. [7] J. Peg ad Z. L, A o-teror cotuato method for geeralzed lear complemetarty problems, Mathematcal Programmg, , pp [8] J. Peg, C. Roos, ad T. Terlaky, Self-Regularty: A ew paradgm for prmal-dual terorpot algorthms, Prceto Uversty Press, [9] R. A. Polyak, Smooth optmzato method for mmax problems, SIAM J. Cotrol ad Optmzato, , pp [20] R.T. Rockafellar, Covex Aalyss, Prceto Uversty Press, Prceto, New Jersey, 970. [2] C.E. Shao, A Mathematcal Theory of Commucato, Bell System Techcal Joural, , pp ad [22] X. L. Su ad D. L, Value-estmato fucto method for costraed global optmzato, Joural of Optmzato Theory ad Applcatos, , pp [23] X. L. Su ad D. L, Logarthmc-expoetal pealty formulato for teger programmg, Appled Mathematcs Letters, 2999, pp [24] X. L. Su ad D. L, Asymptotc strog dualty for bouded teger programmg: A logarthmc-expoetal dual formulato, Mathematcs of Operatos Research, , pp [25] I. Zag, A smoothg techque for m-max optmzato, Mathematcal Programmg, 9 980, pp

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

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

CHAPTER 4 RADICAL EXPRESSIONS

CHAPTER 4 RADICAL EXPRESSIONS 6 CHAPTER RADICAL EXPRESSIONS. The th Root of a Real Number A real umber a s called the th root of a real umber b f Thus, for example: s a square root of sce. s also a square root of sce ( ). s a cube

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

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

Lecture 07: Poles and Zeros

Lecture 07: Poles and Zeros Lecture 07: Poles ad Zeros Defto of poles ad zeros The trasfer fucto provdes a bass for determg mportat system respose characterstcs wthout solvg the complete dfferetal equato. As defed, the trasfer fucto

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

Chapter 9 Jordan Block Matrices

Chapter 9 Jordan Block Matrices Chapter 9 Jorda Block atrces I ths chapter we wll solve the followg problem. Gve a lear operator T fd a bass R of F such that the matrx R (T) s as smple as possble. f course smple s a matter of taste.

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

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

Some Notes on the Probability Space of Statistical Surveys

Some Notes on the Probability Space of Statistical Surveys Metodološk zvezk, Vol. 7, No., 200, 7-2 ome Notes o the Probablty pace of tatstcal urveys George Petrakos Abstract Ths paper troduces a formal presetato of samplg process usg prcples ad cocepts from Probablty

More information

LINEAR RECURRENT SEQUENCES AND POWERS OF A SQUARE MATRIX

LINEAR RECURRENT SEQUENCES AND POWERS OF A SQUARE MATRIX INTEGERS: ELECTRONIC JOURNAL OF COMBINATORIAL NUMBER THEORY 6 2006, #A12 LINEAR RECURRENT SEQUENCES AND POWERS OF A SQUARE MATRIX Hacèe Belbachr 1 USTHB, Departmet of Mathematcs, POBox 32 El Ala, 16111,

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

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

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

5 Short Proofs of Simplified Stirling s Approximation

5 Short Proofs of Simplified Stirling s Approximation 5 Short Proofs of Smplfed Strlg s Approxmato Ofr Gorodetsky, drtymaths.wordpress.com Jue, 20 0 Itroducto Strlg s approxmato s the followg (somewhat surprsg) approxmato of the factoral,, usg elemetary fuctos:

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

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

The internal structure of natural numbers, one method for the definition of large prime numbers, and a factorization test

The internal structure of natural numbers, one method for the definition of large prime numbers, and a factorization test Fal verso The teral structure of atural umbers oe method for the defto of large prme umbers ad a factorzato test Emmaul Maousos APM Isttute for the Advacemet of Physcs ad Mathematcs 3 Poulou str. 53 Athes

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

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

{ }{ ( )} (, ) = ( ) ( ) ( ) 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

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

h-analogue of Fibonacci Numbers

h-analogue of Fibonacci Numbers h-aalogue of Fboacc Numbers arxv:090.0038v [math-ph 30 Sep 009 H.B. Beaoum Prce Mohammad Uversty, Al-Khobar 395, Saud Araba Abstract I ths paper, we troduce the h-aalogue of Fboacc umbers for o-commutatve

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

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

Beam Warming Second-Order Upwind Method

Beam Warming Second-Order Upwind Method Beam Warmg Secod-Order Upwd Method Petr Valeta Jauary 6, 015 Ths documet s a part of the assessmet work for the subject 1DRP Dfferetal Equatos o Computer lectured o FNSPE CTU Prague. Abstract Ths documet

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

4 Inner Product Spaces

4 Inner Product Spaces 11.MH1 LINEAR ALGEBRA Summary Notes 4 Ier Product Spaces Ier product s the abstracto to geeral vector spaces of the famlar dea of the scalar product of two vectors or 3. I what follows, keep these key

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

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

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

Rademacher Complexity. Examples

Rademacher Complexity. Examples Algorthmc Foudatos of Learg Lecture 3 Rademacher Complexty. Examples Lecturer: Patrck Rebesch Verso: October 16th 018 3.1 Itroducto I the last lecture we troduced the oto of Rademacher complexty ad showed

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

EVALUATION OF FUNCTIONAL INTEGRALS BY MEANS OF A SERIES AND THE METHOD OF BOREL TRANSFORM

EVALUATION OF FUNCTIONAL INTEGRALS BY MEANS OF A SERIES AND THE METHOD OF BOREL TRANSFORM EVALUATION OF FUNCTIONAL INTEGRALS BY MEANS OF A SERIES AND THE METHOD OF BOREL TRANSFORM Jose Javer Garca Moreta Ph. D research studet at the UPV/EHU (Uversty of Basque coutry) Departmet of Theoretcal

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

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

TESTS BASED ON MAXIMUM LIKELIHOOD

TESTS BASED ON MAXIMUM LIKELIHOOD ESE 5 Toy E. Smth. The Basc Example. TESTS BASED ON MAXIMUM LIKELIHOOD To llustrate the propertes of maxmum lkelhood estmates ad tests, we cosder the smplest possble case of estmatg the mea of the ormal

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

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

Introduction to local (nonparametric) density estimation. methods

Introduction to local (nonparametric) density estimation. methods Itroducto to local (oparametrc) desty estmato methods A slecture by Yu Lu for ECE 66 Sprg 014 1. Itroducto Ths slecture troduces two local desty estmato methods whch are Parze desty estmato ad k-earest

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

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

DIFFERENTIAL GEOMETRIC APPROACH TO HAMILTONIAN MECHANICS

DIFFERENTIAL GEOMETRIC APPROACH TO HAMILTONIAN MECHANICS DIFFERENTIAL GEOMETRIC APPROACH TO HAMILTONIAN MECHANICS Course Project: Classcal Mechacs (PHY 40) Suja Dabholkar (Y430) Sul Yeshwath (Y444). Itroducto Hamltoa mechacs s geometry phase space. It deals

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

Chapter 2 - Free Vibration of Multi-Degree-of-Freedom Systems - II

Chapter 2 - Free Vibration of Multi-Degree-of-Freedom Systems - II CEE49b Chapter - Free Vbrato of Mult-Degree-of-Freedom Systems - II We ca obta a approxmate soluto to the fudametal atural frequecy through a approxmate formula developed usg eergy prcples by Lord Raylegh

More information

Decomposition of Hadamard Matrices

Decomposition of Hadamard Matrices Chapter 7 Decomposto of Hadamard Matrces We hae see Chapter that Hadamard s orgal costructo of Hadamard matrces states that the Kroecer product of Hadamard matrces of orders m ad s a Hadamard matrx of

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

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

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

Point Estimation: definition of estimators

Point Estimation: definition of estimators Pot Estmato: defto of estmators Pot estmator: ay fucto W (X,..., X ) of a data sample. The exercse of pot estmato s to use partcular fuctos of the data order to estmate certa ukow populato parameters.

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

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

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

. 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

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

Chapter 4 Multiple Random Variables

Chapter 4 Multiple Random Variables Revew for the prevous lecture: Theorems ad Examples: How to obta the pmf (pdf) of U = g (, Y) ad V = g (, Y) Chapter 4 Multple Radom Varables Chapter 44 Herarchcal Models ad Mxture Dstrbutos Examples:

More information

A COMPARATIVE STUDY OF THE METHODS OF SOLVING NON-LINEAR PROGRAMMING PROBLEM

A COMPARATIVE STUDY OF THE METHODS OF SOLVING NON-LINEAR PROGRAMMING PROBLEM DAODIL INTERNATIONAL UNIVERSITY JOURNAL O SCIENCE AND TECHNOLOGY, VOLUME, ISSUE, JANUARY 9 A COMPARATIVE STUDY O THE METHODS O SOLVING NON-LINEAR PROGRAMMING PROBLEM Bmal Chadra Das Departmet of Tetle

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

Multivariate Transformation of Variables and Maximum Likelihood Estimation

Multivariate Transformation of Variables and Maximum Likelihood Estimation Marquette Uversty Multvarate Trasformato of Varables ad Maxmum Lkelhood Estmato Dael B. Rowe, Ph.D. Assocate Professor Departmet of Mathematcs, Statstcs, ad Computer Scece Copyrght 03 by Marquette Uversty

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

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

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

An Introduction to. Support Vector Machine

An Introduction to. Support Vector Machine A Itroducto to Support Vector Mache Support Vector Mache (SVM) A classfer derved from statstcal learg theory by Vapk, et al. 99 SVM became famous whe, usg mages as put, t gave accuracy comparable to eural-etwork

More information

COMPROMISE HYPERSPHERE FOR STOCHASTIC DOMINANCE MODEL

COMPROMISE HYPERSPHERE FOR STOCHASTIC DOMINANCE MODEL Sebasta Starz COMPROMISE HYPERSPHERE FOR STOCHASTIC DOMINANCE MODEL Abstract The am of the work s to preset a method of rakg a fte set of dscrete radom varables. The proposed method s based o two approaches:

More information

NP!= P. By Liu Ran. Table of Contents. The P versus NP problem is a major unsolved problem in computer

NP!= P. By Liu Ran. Table of Contents. The P versus NP problem is a major unsolved problem in computer NP!= P By Lu Ra Table of Cotets. Itroduce 2. Prelmary theorem 3. Proof 4. Expla 5. Cocluso. Itroduce The P versus NP problem s a major usolved problem computer scece. Iformally, t asks whether a computer

More information

STRONG CONSISTENCY OF LEAST SQUARES ESTIMATE IN MULTIPLE REGRESSION WHEN THE ERROR VARIANCE IS INFINITE

STRONG CONSISTENCY OF LEAST SQUARES ESTIMATE IN MULTIPLE REGRESSION WHEN THE ERROR VARIANCE IS INFINITE Statstca Sca 9(1999), 289-296 STRONG CONSISTENCY OF LEAST SQUARES ESTIMATE IN MULTIPLE REGRESSION WHEN THE ERROR VARIANCE IS INFINITE J Mgzhog ad Che Xru GuZhou Natoal College ad Graduate School, Chese

More information

A New Method for Solving Fuzzy Linear. Programming by Solving Linear Programming

A New Method for Solving Fuzzy Linear. Programming by Solving Linear Programming ppled Matheatcal Sceces Vol 008 o 50 7-80 New Method for Solvg Fuzzy Lear Prograg by Solvg Lear Prograg S H Nasser a Departet of Matheatcs Faculty of Basc Sceces Mazadara Uversty Babolsar Ira b The Research

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

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

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

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

ECONOMETRIC THEORY. MODULE VIII Lecture - 26 Heteroskedasticity

ECONOMETRIC THEORY. MODULE VIII Lecture - 26 Heteroskedasticity ECONOMETRIC THEORY MODULE VIII Lecture - 6 Heteroskedastcty Dr. Shalabh Departmet of Mathematcs ad Statstcs Ida Isttute of Techology Kapur . Breusch Paga test Ths test ca be appled whe the replcated data

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

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions.

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions. Ordary Least Squares egresso. Smple egresso. Algebra ad Assumptos. I ths part of the course we are gog to study a techque for aalysg the lear relatoshp betwee two varables Y ad X. We have pars of observatos

More information

A New Measure of Probabilistic Entropy. and its Properties

A New Measure of Probabilistic Entropy. and its Properties Appled Mathematcal Sceces, Vol. 4, 200, o. 28, 387-394 A New Measure of Probablstc Etropy ad ts Propertes Rajeesh Kumar Departmet of Mathematcs Kurukshetra Uversty Kurukshetra, Ida rajeesh_kuk@redffmal.com

More information

The Necessarily Efficient Point Method for Interval Molp Problems

The Necessarily Efficient Point Method for Interval Molp Problems ISS 6-69 Eglad K Joural of Iformato ad omputg Scece Vol. o. 9 pp. - The ecessarly Effcet Pot Method for Iterval Molp Problems Hassa Mshmast eh ad Marzeh Alezhad + Mathematcs Departmet versty of Ssta ad

More information

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions Iteratoal Joural of Computatoal Egeerg Research Vol, 0 Issue, Estmato of Stress- Stregth Relablty model usg fte mxture of expoetal dstrbutos K.Sadhya, T.S.Umamaheswar Departmet of Mathematcs, Lal Bhadur

More information

Can we take the Mysticism Out of the Pearson Coefficient of Linear Correlation?

Can we take the Mysticism Out of the Pearson Coefficient of Linear Correlation? Ca we tae the Mstcsm Out of the Pearso Coeffcet of Lear Correlato? Itroducto As the ttle of ths tutoral dcates, our purpose s to egeder a clear uderstadg of the Pearso coeffcet of lear correlato studets

More information

STRONG CONSISTENCY FOR SIMPLE LINEAR EV MODEL WITH v/ -MIXING

STRONG CONSISTENCY FOR SIMPLE LINEAR EV MODEL WITH v/ -MIXING Joural of tatstcs: Advaces Theory ad Alcatos Volume 5, Number, 6, Pages 3- Avalable at htt://scetfcadvaces.co. DOI: htt://d.do.org/.864/jsata_7678 TRONG CONITENCY FOR IMPLE LINEAR EV MODEL WITH v/ -MIXING

More information

CS286.2 Lecture 4: Dinur s Proof of the PCP Theorem

CS286.2 Lecture 4: Dinur s Proof of the PCP Theorem CS86. Lecture 4: Dur s Proof of the PCP Theorem Scrbe: Thom Bohdaowcz Prevously, we have prove a weak verso of the PCP theorem: NP PCP 1,1/ (r = poly, q = O(1)). Wth ths result we have the desred costat

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

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

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

Chapter 4 (Part 1): Non-Parametric Classification (Sections ) Pattern Classification 4.3) Announcements

Chapter 4 (Part 1): Non-Parametric Classification (Sections ) Pattern Classification 4.3) Announcements Aoucemets No-Parametrc Desty Estmato Techques HW assged Most of ths lecture was o the blacboard. These sldes cover the same materal as preseted DHS Bometrcs CSE 90-a Lecture 7 CSE90a Fall 06 CSE90a Fall

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

Investigating Cellular Automata

Investigating Cellular Automata Researcher: Taylor Dupuy Advsor: Aaro Wootto Semester: Fall 4 Ivestgatg Cellular Automata A Overvew of Cellular Automata: Cellular Automata are smple computer programs that geerate rows of black ad whte

More information

Sufficiency in Blackwell s theorem

Sufficiency in Blackwell s theorem Mathematcal Socal Sceces 46 (23) 21 25 www.elsever.com/locate/ecobase Suffcecy Blacwell s theorem Agesza Belsa-Kwapsz* Departmet of Agrcultural Ecoomcs ad Ecoomcs, Motaa State Uversty, Bozema, MT 59717,

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

Research Article Gauss-Lobatto Formulae and Extremal Problems

Research Article Gauss-Lobatto Formulae and Extremal Problems Hdaw Publshg Corporato Joural of Iequaltes ad Applcatos Volume 2008 Artcle ID 624989 0 pages do:055/2008/624989 Research Artcle Gauss-Lobatto Formulae ad Extremal Problems wth Polyomals Aa Mara Acu ad

More information

Third handout: On the Gini Index

Third handout: On the Gini Index Thrd hadout: O the dex Corrado, a tala statstca, proposed (, 9, 96) to measure absolute equalt va the mea dfferece whch s defed as ( / ) where refers to the total umber of dvduals socet. Assume that. The

More information

1 Convergence of the Arnoldi method for eigenvalue problems

1 Convergence of the Arnoldi method for eigenvalue problems Lecture otes umercal lear algebra Arold method covergece Covergece of the Arold method for egevalue problems Recall that, uless t breaks dow, k steps of the Arold method geerates a orthogoal bass of a

More information

Research Article Some Strong Limit Theorems for Weighted Product Sums of ρ-mixing Sequences of Random Variables

Research Article Some Strong Limit Theorems for Weighted Product Sums of ρ-mixing Sequences of Random Variables Hdaw Publshg Corporato Joural of Iequaltes ad Applcatos Volume 2009, Artcle ID 174768, 10 pages do:10.1155/2009/174768 Research Artcle Some Strog Lmt Theorems for Weghted Product Sums of ρ-mxg Sequeces

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

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

L5 Polynomial / Spline Curves

L5 Polynomial / Spline Curves L5 Polyomal / Sple Curves Cotets Coc sectos Polyomal Curves Hermte Curves Bezer Curves B-Sples No-Uform Ratoal B-Sples (NURBS) Mapulato ad Represetato of Curves Types of Curve Equatos Implct: Descrbe a

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

NP!= P. By Liu Ran. Table of Contents. The P vs. NP problem is a major unsolved problem in computer

NP!= P. By Liu Ran. Table of Contents. The P vs. NP problem is a major unsolved problem in computer NP!= P By Lu Ra Table of Cotets. Itroduce 2. Strategy 3. Prelmary theorem 4. Proof 5. Expla 6. Cocluso. Itroduce The P vs. NP problem s a major usolved problem computer scece. Iformally, t asks whether

More information

Lecture 3. Sampling, sampling distributions, and parameter estimation

Lecture 3. Sampling, sampling distributions, and parameter estimation Lecture 3 Samplg, samplg dstrbutos, ad parameter estmato Samplg Defto Populato s defed as the collecto of all the possble observatos of terest. The collecto of observatos we take from the populato s called

More information