Probabilistic Engineering Mechanics. Do Rosenblatt and Nataf isoprobabilistic transformations really differ?
|
|
- Gwenda Randall
- 8 years ago
- Views:
Transcription
1 Probabilistic Egieerig Mechaics 4 (009) Cotets lists available at ScieceDirect Probabilistic Egieerig Mechaics joural homepage: wwwelseviercom/locate/probegmech Do Roseblatt ad Nataf isoprobabilistic trasformatios really differ? Régis Lebru a,, Ae Dutfoy b a EADS Iovatio Wors, System Egieerig - Applied Mathematics, Frace b EDF R&D, Idustrial Ris Maagemet, Frace a r t i c l e i f o a b s t r a c t Article history: Received 9 September 008 Received i revised form 9 February 009 Accepted 3 April 009 Available olie April 009 Keywords: Roseblatt trasformatio Elliptical copula Nataf trasformatio This article is the third i a series dedicated to the mathematical study of isoprobabilistic trasformatios ad their relatioship with stochastic depedece modellig, see [R Lebru, A Dutfoy, A iovatig aalysis of the Nataf trasformatio from the viewpoit of copula, Probabilistic Egieerig Mechaics (008) doi: 006/jprobegmech ] for a iterpretatio of the Nataf trasformatio i term of ormal copula ad [R Lebru, A Dutfoy, A geeralizatio of the Nataf trasformatio to distributios with elliptical copula, Probabilistic Egieerig Mechaics (4) (009), 7 78 doi:006/jprobegmech ] for a geeralisatio of the Nataf trasformatio to ay elliptical copula I this article, we explore the relatioship betwee two isoprobabilistic trasformatios widely used i the commuity of reliability aalysts, amely the Geeralised Nataf trasformatio ad Roseblatt trasformatio First, we recall the elemetary results of the copula theory that are eeded i the remaiig of the article, as a prelimiary sectio to the presetatio of both the Geeralized Nataf trasformatio ad the Roseblatt trasformatio i the light of the copula theory The, we show that the Roseblatt trasformatio usig the caoical order of coditioig is idetical to the Geeralised Nataf trasformatio i the ormal copula case, which is the most usual case i reliability aalysis sice it correspods to the classical Nataf trasformatio At this step, we also show that it is ot possible to exted the Roseblatt trasformatio to distributios with geeral elliptical copula the way the Nataf trasformatio has bee geeralised Furthermore, we explore the effect of the coditioig order of the Roseblatt trasformatio o the usual reliability idicators obtaied from a FORM or SORM method We show that i the ormal copula case, all these reliability idicators, excepted the importace factors, are uchaged whatever the coditioig order oe chooses I the last sectio, we coclude the article with two umerical applicatios that illustrate the previous results: the equivalece betwee both trasformatios i the ormal copula case, ad the effect of the coditioig order i the ormal ad o-ormal copula case 009 Published by Elsevier Ltd Itroductio I reliability aalysis, the ucertaity related to the iput variables is modelled with a multivariate probability distributio that defies a radom vector X of dimesio We suppose i this article that this distributio is absolutely cotiuous ad thus admit a joit desity fuctio These distributios are propagated through a physical model f to compute the distributio of the output variable Y Staeholders aim at taig a decisio o the Correspodig address: EADS Iovatio Wors, System Egieerig - Applied Mathematics, rue Pasteur, BP76, 95 Sureses, Frace Tel: ; fax: addresses: regislebru@eadset, regislebru@m4xorg (R Lebru), aedutfoy@edffr (A Dutfoy) basis of the realisatio of criteria, which may be the probability that the output variable overpasses a give threshold The evaluatio of this probability might require a high umber of simulatios, which might forbid the use of simulatio algorithms, particularly if the physical model has a high computatioal cost That is why a alterative to the simulatio methods has bee developed, based o a isoprobabilistic trasformatio that maps the physical space ito a ew space called the stadard space I that space, the evaluatio of the probability ca be doe usig cheap approximatios such as the FORM or SORM approximatios Two isoprobabilistic trasformatios are preseted i the literature: the Nataf trasformatio [,] which has bee exteded ito a Geeralised Nataf trasformatio [] ad the Roseblatt trasformatio [3] /$ see frot matter 009 Published by Elsevier Ltd doi:006/jprobegmech
2 578 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) The mai objective of this article is to compare the Geeralized Nataf trasformatio with the Roseblatt oe ad to prove that they are idetical i the ormal copula case, which is the most commo case i actual reliability studies as it correspods to the use of the classical Nataf trasformatio We also study the possibility to modify the Roseblatt trasformatio i a similar way that the Nataf trasformatio has bee exteded to lead to a o-ormal stadard space The secod objective of this article is to study the impact of the coditioig order of the Roseblatt trasformatio o the usual reliability idicators obtaied after a aalytical FORM/SORM method, with a focus o the ormal copula case We ote F X the -th uivariate margial cumulative distributio fuctio of the radom vector X ad C X its copula We ote F X, the cumulative distributio fuctio of the sub-vector (X,, X ) of X ad C X, its copula Whe o cofusio is possible, we remove the superscript i order to ease the readig: F X ad C X become F ad C Furthermore, we ote C N R is a ormal copula whose correlatio matrix is R We suppose that R is a symmetric positive defiite matrix We ote M, (R) the set of a real square matrix of dimesio, O (R) the subset of orthogoal matrices of M, (R), ie: Q O (R), Q Q t = I () ad GL (R) the subset of ivertible matrices of M, (R) If R = (r ij ) i,j M, (R), the R is its -leadig sub-bloc: R = (r ij ) i,j () ad R is the ( + )-th partial colum vector: R = (r,+,, r,+ ) t (3) We call stadard space the image space of a isoprobabilistic trasformatio Elemetary results o copula I this sectio, we recall some basic results of the theory of copulas We restrict ourselves to the strict miimum to follow further developmets The iterested reader is ivited to cosult eg [4] for a thorough itroductio to the theory of copulas, icludig a demostratio of the results preseted here The followig theorem lis the joit cumulative distributio fuctio of a multivariate distributio to a copula: Theorem (Slar, 959) Let F be a cumulative distributio fuctio of dimesio whose uivariate margial cumulative distributio fuctios are F i It exists a copula C of dimesio such that for x R, we have: F(x,, x ) = C(F (x ),, F (x )) (4) If the margial distributios F i are cotiuous, the copula C is uique; otherwise, it is uiquely determied o Rage(F ) Rage(F ) I the case of cotiuous margial distributios, for all u [0, ], we have: C(u) = F ( F (u ),, F (u ) ) (5) The copula is uchaged uder almost strictly icreasig margial trasformatio: Propositio If X has as a copula C ad if (α,, α ) are almost strictly icreasig fuctios defied respectively o the supports of the X i, the C is also the copula of (α (X ),, α (X )) If we are iterested i the distributio of the sub-vector (X,, X ) of the radom vector X, we have the followig property: Propositio 3 (-dimesioal Margial Distributios) Let X be a cotiuous radom vector with a distributio defied by its copula C ad its margial cumulative distributio fuctios F i The cumulative distributio fuctio F, of the -dimesioal radom vector (X,, X ) is defied by its margial distributios F i ad the copula C, through the relatio: F, (x,, x ) = C, (F (x ),, F (x )) (6) with C, (u,, u ) = C(u,, u,,, ) (7) Defiitio 4 (-th Coditioal Margial Distributio) Let X = (X,, X ) be a cotiuous radom vector The cumulative distributio fuctio of the coditioal variable X X,, X is defied by: F,, (x x,, x ) = F, (x,, x ) x x / F, (x,, x ) x x (8) Propositio 5 (-th Coditioal Margial Copula) Let X be a cotiuous radom vector with a distributio defied by its copula C ad its margial cumulative distributio fuctios F i The cumulative distributio fuctio of the coditioal variable X X,, X is defied by its margial distributios F i ad the copula C,, through the relatio: F,, (x x,, x ) = C,, (F (x ) F (x ),, F (x )) (9) with C,, (u u,, u ) = C, (u,, u ) u u / C, (u,, u ) u u (0) As a matter of fact, relatio (0) is the direct applicatio of Defiitio 4 to the cumulative distributio C Furthermore, Defiitio 4 ad relatio (4) lead to: F,, (x x,, x ) [ ] i= / C, (F (x ),, F (x )) = p i (x i ) u i= u [ ] i= C, (F (x ),, F (x )) p i (x i ) () u u i= where p i is the probability desity fuctio of X i Relatios () ad (0) lead to relatio (9) 3 The Geeralised Nataf ad Roseblatt trasformatios The Nataf trasformatio [] has bee itroduced as a procedure to trasform the uivariate margial distributios of a multivariate margial distributio Its usage i reliability aalysis has bee popularised by several authors, see eg [5,6] By the way, the Nataf trasformatio has bee preseted as
3 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) a way to perform reliability computatios usig icomplete depedece iformatio, ad it is oly recetly that its role as a probabilistic modellig tool has bee emphasised, lied to the copula theory [7]: the traditioal usage of the Nataf trasformatio supposes that the radom vector X has a ormal copula A geeralisatio to radom vectors with ay elliptical copulas has bee proposed i [], ad this is the geeralisatio which is recalled here: Defiitio 6 (Geeralised Nataf Trasformatio) Let X i R be a cotiuous radom vector defied by its uivariate margial cumulative distributio fuctios F X i ad its elliptical copula C X, characterized by its geerator ψ The Geeralised Nataf trasformatio T GN is defied by: U = T GN (X) = T GN T GN (X) () where both trasformatios T GN ad T GN are give by: T GN : R R X W = E F X (X ) E F X (X ) T GN : R R W U = Γ W (3) where E is the cumulative distributio fuctio of uivariate stadard elliptical distributio with characteristic geerator ψ ad Γ is the iverse of the Cholesy factor of R Aother widely used isoprobabilistic trasformatio is the Roseblatt trasformatio [3], defied by: Defiitio 7 (Roseblatt Trasformatio) Let X i R be a cotiuous radom vector defied by its uivariate margial cumulative distributio fuctios F X i ad its copula C X The Roseblatt trasformatio T R of X is defied by: U = T R (X) = T R T R (X) (4) where both trasformatios T R, ad T R are give by: T R : R R F X (X ) X Y = F X,, (X X,, X ) F X,, (X X,, X ) T R : R R Φ (Y ) Y U = Φ (Y ) (5) where F X,, is the cumulative distributio fuctio of the coditioal radom variable X X,, X Let us ote that T R maps X ito a uiformly distributed radom vector over [0, ] with idepedet copula For the demostratio, see [5, chap 7, p 3] Furthermore, T R maps Y ito a ormally distributed radom vector with zero mea ad uit covariace matrix: as a mater of fact, Propositio implies that U has the same copula as Y Besides, by costructio, the uivariate margial distributios of U are stadard ormal I order to ease the further compariso betwee the geeralised Nataf trasformatio ad the Roseblatt oe, it is useful to rewrite the Roseblatt trasformatio as follows: Defiitio 8 (Roseblatt Trasformatio, New Formulatio) Let X i R be a cotiuous radom vector defied by its uivariate margial cumulative distributio fuctios F X i ad its copula C X The ew formulatio of the Roseblatt trasformatio T NR is defied by: U = T NR (X) = T R T 0 (X) (6) where T 0 is give by: T 0 : R R X W = Φ F X (X ) (7) Φ F X (X ) where Φ is the cumulative distributio fuctio of the uivariate stadard ormal distributio, T R the Roseblatt trasformatio of Defiitio 7 Let us ote that U NR = T NR (X) = T R T R T 0(X) If W = T 0 (X), the, thas to (5), the th compoet of U NR writes: U NR = Φ F W,, (W W,, W ) (8) Thas to Propositio 5, the cumulative distributio fuctio of the coditioal variable W W,, W writes: F W (w,, w,, w ) ( ) = C W,, F W (w ) F W (w ),, F W (w ) (9) From Propositio, it follows that X ad W have the same copula C X Furthermore, by costructio of W, we have F W = Φ ad Φ(W ) = F X (X ) The, relatio (9) rewrites: F W,, (W W,, W ) ( ) = C X,, F X (X ) F X (X ),, F X (X ) which fially leads to the relatio: (0) F W,, (W W,, W ) = F X,, (X X,, X ) () ad the to: U NR = Φ F X,, (X X,, X ) () which is precisely the expressio of the Roseblatt trasformatio of Defiitio 7: U NR = T R (X) 4 Do geeralised Nataf ad Roseblatt trasformatios really differ? I this sectio, we first cosider the case where the copula of X is ormal, which is the most usual case i reliability aalysis sice it correspods to the case where the classical Nataf trasformatio applies The, we mae the compariso i all the other cases: o-ormal elliptical copulas ad o-elliptical copulas 4 The ormal copula case The ew formulatio (6) of the Roseblatt trasformatio maes it easier to show that whe X has a ormal copula, both trasformatios are idetical:
4 580 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) Propositio 9 (Isoprobabilistic Trasformatios, Normal Copula) Let X i R be a cotiuous radom vector defied by its uivariate margial cumulative distributio fuctios F X i ad its ormal copula The, the Roseblatt trasformatio ad the geeralised Nataf oe C N R are idetical: T R (X) = T GN (X) (3) We recall without demostratio some well-ow results about ormal vectors ad a easy-to demostrate result o orthogoal matrices that will be used i the demostratio of Propositio 9 Propositio 0 (Coditioal Normal Vector) Let U = (X, X ) t i R + be a ormal radom vector The the coditioal radom vector V = X X is a ormal vector which mea vector ad covariace matrix are defied by: { E[V] = E[X ] + Cov (X, X )[Cov (X, X )] (X E[X ]) Cov (V) = Cov (X, X ) Cov (X, X )[Cov (X, X )] Cov (X, X ) (4) Propositio (Affie Trasformatio) Let X i R be a ormal vector, with mea vector is µ, ad covariace matrix Σ, A a determiistic matrix i M,p (R) ad b i R p a determiistic vector The Y = A X + b is a ormal vector whose mea vector ad covariace matrix are defied by: { E[Y] = A µ + b Cov (Y) = A Σ A t (5) Propositio (Orthogoal Triagular Matrix with Positive Diagoal) Let T + (R) be the set of lower triagular matrix of M, (R) with positive diagoal elemets The T + (R) is a multiplicative subgroup of GL (R) Furthermore, T + (R) O (R) = {I } We ca ow start to demostrate Propositio 9, usig the ew formulatio of the Roseblatt trasformatio of Defiitio 8, whose differet steps are the followig oes: T NR : X T 0 W T R Y T R U (6) Let us ote S = (W,, W ) t ad V = W S As X has a ormal copula, W is a -dimesioal ormal vector whose uivariate margial distributios are stadard ormal ad which correlatio matrix is R Propositio 0 gives that for all, V follows a uivariate ormal distributio ad relatio (4) leads to: E[V ] = E[W ] + Cov (W, S )[Cov (S, S )] (S E[S ]) = Cov (W, S )[Cov (S, S )] S (7) Besides, we have Cov (W) = Cor (W) = R, ad give the otatios () ad (3), we have: E[V ] = ( R ) t [R ] S (8) Furthermore, relatio (4) also leads to: Var[V ] = Var(W ) Cov (W, S )[Cov (S, S )] Cov (S, W ) = ( R ) t [R ] R (9) Give relatios (8) ad (9), the th compoet of Y is defied by: Y = F W,, (W W,, W ) = Φ W ( R )t [R ] S ( (30) R )t [R ] R Fially, we obtai: U = Φ (Y ) = W ( R )t [R ] S ( R )t [R ] R = A W (3) where for all [, ], A = (a,,, a,, 0,, 0) M (R) with: a, = [ ( ] R )t [R ] R i= (3) a,j = a, r i r ji for j [, ] i= As A is a row matrix, U oly depeds o S Let Γ be the lower triagular matrix which lie is A The relatio (3) implies that: U = Γ W (33) which is very close to relatio (3) It remais to show that Γ = Γ Propositio implies that Cov (U) = Γ R Γ t ad Cov (U) = I by costructio of U If L is the Cholesy factor of R, the R = L L t, ad ( Γ L )( Γ L ) t = I, which leads to Γ L O (R) Furthermore, by costructio, Γ T + (R) As L T + (R), Propositio implies that Γ L T + (R) ad Γ L = I, which rewrites Γ = L = Γ I coclusio, we showed that i the case where X has a ormal copula, we have the relatio T R T R T 0(X) = T N T N (X) which leads to: T R (X) = T N (X) (34) Thus, the equivalece of the Roseblatt trasformatio ad the Geeralised Nataf trasformatio i the ormal copula case is show 4 The other cases I the case where the copula of X is elliptical but o-ormal, both isoprobabilistic trasformatios differ as their associated stadard spaces are differet As a matter of fact, the stadard spaces of the Geeralised Nataf is associated with the stadard spherical represetative of the elliptical family that defies the elliptical copula, whereas the stadard space of the Roseblatt trasformatio is associated to the ormal distributio At this step, it is iterestig to chec whether it is possible to modify the Roseblatt trasformatio i order to mae its stadard space be the same as the oe associated with the geeralised Nataf trasformatio I [7] we recall that the essetial characteristic of the stadard space is the spherical symmetry of its associated distributio, which gives a sese to the FORM ad SORM approximatios of the evet probability Let us ote that by costructio, because of the coditioig step T R, the Roseblatt trasformatio leads to a fial vector U with a idepedet copula
5 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) orthogoal trasformatio i the stadard space of the Roseblatt trasformatio More precisely, if we ote P the permutatio matrix associated to the arbitrary order, T R the Roseblatt trasformatio associated to this orderig, U = T R(X) ad U = T R (X), the we have: Q O (R) / U = Q U (35) Fig Roseblatt trasformatios whe the coditioig of the compoets W follow the caoical order or a arbitrary order Propositio 3 (Spherical Distributio with Idepedet Copula) The oly spherical distributios with idepedet compoets are the ormal distributios with zero mea ad scalar covariace matrix λi with λ > 0 See [8] for a demostratio Thus, the oly way to map a radom vector with a idepedet copula ito a radom vector followig a spherical distributio, is to map it ito a ormal vector such as described i this propositio: thus, the stadard space of the Roseblatt trasformatio is ecessarily the ormal oe Therefore, the stadard space of the Roseblatt trasformatio ad the stadard space of the Geeralised Nataf trasformatio oly coicide i the ormal copula case At last, for all the other cases where the copula of X is ot elliptical, the Geeralised Nataf trasformatio is ot defied ad the compariso with the Roseblatt trasformatio ot possible 5 Impact of the coditioig order i the Roseblatt trasformatio i the ormal copula case I the literature [5], the presetatio of the Roseblatt trasformatio is give with the warig that the coditioig order i step T R has a impact o the results obtaied from a FORM/SORM method Let us call caoical order the order preseted i the relatio (5) I that sectio, we study the impact of a chage i the coditioig order of the Roseblatt trasformatio o the quatities evaluated i the cotext of the use of the FORM or SORM methods: the desig poit, which is used through its orm (reliability idex) ad its ormalised squared coordiates (importace factors), ad the curvatures of the limit state surface at the desig poit i the stadard space, where the limit state surface is the frotier of the subspace of parameters verifyig the evet (for SORM approximatio) I the case where the copula of X is ot ormal, it has already bee show that such a chage has a impact o all these elemets: see the example quoted by [9] ad discussed by several authors, for example [5] ad [0] However, this is ot always the case We will study i more detail the most frequet situatio where the copula of X is ormal sice, as metioed previously, it is the copula iduced by the use of the classical Nataf trasformatio Let us suppose ow that we chage the order of coditioig It is equivalet to cosider the itroductio of a ew step i the Roseblatt trasformatio betwee the steps T 0 ad T R of relatio (6) i order to mae a permutatio of the compoets of W We ote P the permutatio matrix such as W = P W The Roseblatt trasformatios usig the caoical order or a arbitrary order are summarised graphically i Fig We have the followig result: Propositio 4 (Impact of the Order of Coditioig, Normal Copula) I the ormal copula case, chagig the order of the coditioig i the Roseblatt trasformatio cosists i maig a where Q ad P are i the same coected compoet of O (R), it meas det P = det Q Accordig to the otatios of Fig, if R is the correlatio matrix of the ormal copula of X, R the oe of W, Γ ad Γ the iverse of their respective Cholesy factors, the the matrices P ad Q are lied by: Q = Γ P Γ (36) The followig result will help for the demostratio of Propositio 4: Propositio 5 (Triagular Decompositio) Let A ad B be two determiistic matrices i M, (R), with B ivertible The we have: A A t = B B t B A O (R) (37) which meas that Q O (R) / A = B Q As a matter of fact, we have the followig implicatios: (B A )(B A ) t = B A A t B t = B B B t B t = I (38) which leads to the result of Propositio 5 As W = P W, W is a ormal vector which correlatio matrix verifies R = P R P t ad which Cholesy factor is L = Γ Therefore, R = L L t = (P L )(P L ) t Propositio 5 leads to: Q O (R) / P L = L Q (39) By multiplyig the relatio (39) o the left by Γ ad o the right by Γ, it rewrites: Γ P = Q Γ (40) which leads to the relatio betwee P ad Q give i relatio (36) We showed that i the ormal copula case, the mappig from W ito U is liear such as: U = Γ W Fially, we obtai: U = Γ P W (4) Relatios (40) ad (4) fially imply that: U = Q Γ W = Q U (4) as required Give that det(γ ) > 0 ad det(γ ) > 0, relatio (36) implies that det(q ) ad det(p ) have the same sig, which sigifies that they belog to the same coected compoet of O (R) I coclusio, if the radom vector X has a ormal copula, the effect of chagig the order of coditioig i the Roseblatt trasformatio with respect to the caoical order is to apply a further orthogoal trasformatio after applyig the Roseblatt trasformatio associated to the caoical orderig It chages the locatio of the desig poit, therefore its coordiates, but either its orm or the curvatures of the limit state surface at the desig poit Thus, i the cotext of the FORM or SORM method, the followig quatities do ot deped o the coditioig order of the Roseblatt trasformatio:
6 58 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) the Hasofer reliability idex [], which is the orm of the desig poit, the FORM approximatio of the evet probability which relies oly o the Hasofer reliability idex, the SORM approximatios of the evet probability which rely o both the Hasofer reliability idex ad the curvatures of the limit state fuctio at the desig poit However, the importace factors which are evaluated from the ormalised squared coordiates of the desig poit chage i a way which is ot i geeral a permutatio of the values obtaied usig the caoical order: relatio (36) implies that i geeral, Q P To be more precise, let us cosider the followig propositio: Propositio 6 (Orderig Idifferece) ( permutatio matrix P, Q = P ) (X has a idepedet copula) The first implicatio is obvious: if X has a idepedet copula, the correlatio matrix R is equal to the idetity matrix I, which implies that R = I, Γ = I, Γ = I ad fially Q = P The secod implicatio derives from the followig computatio By defiitio of Γ ad Γ, we have: Q = P L = P L P t (43) which implies the followig relatio o the coefficiets of L = (l i,j ) i,j ad L = (l i,j ) i,j : l i,j = l σ (i),σ (j) (44) where σ is the permutatio associated to P Thus, give that L ad L are lower triagular matrices, if the relatio (44) must hold for all the permutatios σ, it must hold i particular for ay traspositio τ ij that exchages i ad j, thus if i < j, l i,j = 0 by costructio, thus l σ (i),σ (j) = l ji = 0: L is a diagoal matrix ad cosequetly, R = I, which is equivalet to the idepedece of the compoets of X i the ormal copula case I coclusio, a permutatio with respect to the caoical order o the compoets of X always correspods to the same permutatio with respect to the caoical order of the compoets of the stadard space radom vector oly o the idepedet case Otherwise, the choice of the coditioig order does ot traslate ito a simple permutatio of the values of the desig poit coordiates This remar re-eforces the fact that it is difficult to iterpret the importace factors o a compoet basis i the case of correlated variables Remar 7 A similar questio may be raised cocerig the Nataf trasformatio, amely the choice of Γ, which is by o way restricted to the Cholesy factor for the stadard space to be associated with a spherical distributio: ay square root of the correlatio matrix would suit I particular, if P is a permutatio matrix (ad more geerally ay orthogoal matrix), L P is such a square root Let us recall that the exact value of the evet probability remais uchaged whatever the trasformatio we use, ad whatever the coditioig order we use for the Roseblatt trasformatio! 6 Numerical applicatios I this sectio, we illustrate the results obtaied i the previous sectios through two umerical applicatios We cosider a bi-dimesioal radom vector X = (X, X ) defied by its margial cumulative distributio fuctios (F, F ) ad its copula C For both applicatios, we choose expoetial distributios X Exp(λ ) ad X Exp(λ ) for the margial distributios ad a limit state surface defied by: 8X + X = 0 (45) We cosider the evet: 8X + X 0 (46) which we wat to evaluate the probability I the first applicatio, we choose a ormal copula C N R ( ) where R = ρ ρ ad ρ the correlatio coefficiet We chec both the equivalece betwee the caoical Roseblatt trasformatio ad the Geeralised Nataf trasformatio, ad the effect of a chage i the coditioig order I the secod applicatio, we choose o-elliptical copula, amely the Fra copula C θ, which belogs to the class of archimedea copulas, ad we verify that a chage i the coditioig order is ot equivalet to a orthogoal modificatio of the trasformatio ad has a impact o the FORM ad SORM approximatios We recall that the Fra copula is defied o [0, ] by the expressio: C θ (u, u ) = ( θ log + (e θu )(e θu ) ) e θ (47) where θ R For θ = 0, C θ is the idepedet copula I the umerical applicatios, we tae λ =, λ = 3, ρ = / ad θ = 0 6 Applicatio : Normal copula We use the ew expressio of the Roseblatt trasformatio of Defiitio 8, with the previous otatio W = T 0 (X), i that particular case of ormal copula Give Propositio 0, W W is a ormal radom vector such as E[W W ] = ρw ad( Var[W W ) ] = ρ, which implies that F W W (W W W ) = Φ ρw ρ Fially, the radom vector U is defied by: U = Φ F W (W ) = W U = Φ F W W (W W ) = W ρw ρ (48) The Roseblatt trasformatio with caoical order o the coditioig step fially defies the ormal radom vector U as: U = Φ F (X ) U = Φ F (X ) ρφ F (X ) (49) ρ I the Roseblatt stadard space, the limit state surface has the parametric expressio, where ξ [0, + [: u = Φ F (ξ) ( ) Φ F 8ξ ρφ F (ξ) (50) u = ρ
7 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) Limit State Surface i Stadard Space through Roseblatt Trasformatios Limit State Surface i Stadard Space through Roseblatt trasformatios u Caoical order Desig poit Iverse order Desig poit u Caoical order desig poit Iverse order desig poit u u Fig Trasformatios of the limit state surface ito the stadard space whe usig the caoical order i the Roseblatt trasformatio ad its iverse The liear correlatio is ρ = /, the copula is ormal, X Exp() ad X Exp(3) The limit state surface is 8X + X = 0 Note the symmetry that exchages the two curves: its matrix is Q With the same cosideratios, the Roseblatt trasformatio with the iverse order o the coditioig step defies the ormal radom vector Ũ as: Ũ = Φ F (X ) Ũ = Φ F (X ) ρφ F (X ) (5) ρ which leads, i the stadard space, to the other expressio of the limit state surface: ( ) 8ξ ũ = Φ F ( ) Φ F (ξ) ρφ F 8ξ (5) ũ = ρ Fig draws the graph of the limit state surface i the stadard space after both Roseblatt trasformatios Thas to relatio (36), we ca defie the orthogoal matrix Q The permutatio matrix is P = ( 0 0 Furthermore, we have 0 Γ = Γ = ρ ρ ρ ad fially ( ) ρ ρ Q = ρ ρ ) which leads to R = R We ca easily verify that Ũ = Q U Furthermore, Q is a permutatio matrix with det(q ) =, as the matrix P ( ) +ρ The director vector of the symmetry axis is, ρ I the umerical applicatio draw i Fig, the symmetry axis is ( 3/, /) Fig 3 Trasformatios of the limit state surface ito the stadard space whe usig the caoical order i the Roseblatt trasformatio ad its iverse The copula is a Fra oe with θ = 0, X Exp() ad X Exp(3) The limit state surface is 8X + X = 0 The Hasofer reliability idex is β = 30 ad the FORM approximatio of the evet probability p = P(8X + X < 0) is: p FORM = Φ ( β) = (53) A aalytical computatio of p leads to the umerical result: p = (54) Let us verify ow the equivalece betwee the Nataf trasformatio ad the Roseblatt oe (give that we cosider the caoical order) The Nataf trasformatio leads to the ormal radom vector U defied as: ( ) Φ U = F (X Γ ) Φ F (55) (X ) ( ) As Γ = 0 ρ ρ ρ, we have: U = Φ F (X ) U = ρφ F (X ) + Φ F (X ) (56) ρ ρ which is idetical to the expressio defied i (49) 6 Applicatio : Fra copula We cosider here the Fra copula, which is a o-elliptical copula This example proves that both limit state surfaces i the stadard space associated to two differet orders i the coditioig step of the Roseblatt trasformatio are ot lied by a orthogoal trasformatio We also illustrate that, accordig to this coditioig order, both reliability idex are differet which leads to differet FORM approximatios of the probability Fig 3 draws the graph of the limit state fuctio i the stadard space after both Roseblatt trasformatios The respective reliability idex are differet i both cases: { βcaord = 4 (57) β IvOrd = 7
8 584 R Lebru, A Dutfoy / Probabilistic Egieerig Mechaics 4 (009) which leads to differet FORM approximatios of the evet probability: { P FORM CaOrd = 07 0 P FORM = IvOrd (58) 0 There is a differece of 4% betwee the two approximatios, oly due to the coditioig order, whereas the exact probability value is the same A aalytical computatio of p leads to the umerical value: p = (59) 7 Coclusio This article is the third part of global reflectios o isoprobabilistic trasformatios Its first mai objective was to compare the geeralised Nataf trasformatio with the Roseblatt trasformatio ad show that i the ormal copula case, both trasformatios are idetical I the use of the Roseblatt trasformatio, there is a degree of freedom i the orderig of the coditioig step This poit is ofte preseted as a drawbac of this trasformatio, as it leads to differet umerical results for the FORM ad SORM approximatio The secod mai objective of the article was to show that, although the coditioig order has such a impact i geeral, i the ormal copula case there is ideed o impact o the FORM ad SORM approximatios as well as o the reliability idex The oly impact is o the importace factors i the case of correlated compoets for X, which uderlies the difficulty i iterpretig such factors i the correlated case The Nataf trasformatio has bee successfully geeralised to produce more geeral stadard spaces tha the ormal oe We showed that the Roseblatt trasformatio caot be geeralised this way Thus, for the case of a o-ormal elliptical copula, oe ca choose betwee both isoprobabilistic trasformatios: the Roseblatt trasformatio or the geeralised Nataf trasformatio We illustrated these results through two umerical applicatios, showig the equivalece of both trasformatios i the ormal copula case ad the effect of the coditioig order i a ormal ad o-ormal copula case Let us recall that the exact value of the evet probability remais the same whatever the trasformatio we use, ad whatever the coditioig order we use for the Roseblatt trasformatio It is oly the FORM ad SORM approximatios that are potetially modified Refereces [] Nataf A Détermiatio des distributios de probabilité dot les marges sot doées, Comptes Redus de l Académie des Scieces 5: 4 43 [] Lebru R, Dutfoy A A geeralizatio of the Nataf trasformatio to distributios with elliptical copula Probabilistic Egieerig Mechaics 009; 4:7 8 doi:006/jprobegmech [3] Roseblatt M Remars a multivariate trasformatio The Aals of Mathematical Statistics 3 (3): [4] Nelse R A itroductio to copulas Spriger; 999 [5] Ditlevse O, Madse HO Structural reliability methods Iteret editio March 004 [6] Der Kiureghia A, Liu PL Structural reliability uder icomplete probabilistic iformatio Joural of Egieerig Mechaics 986;():85 04 [7] Lebru R, Dutfoy A A iovatig aalysis of the Nataf trasformatio from the viewpoit of copula Probabilistic Egieerig Mechaics 008 doi:006/jprobegmech [8] Arold SF, Lych J O Ali s characterizatio of the spherical ormal distributio Joural of Royal Statistical Society 98;44():49 5 [9] Dolisy K First-order secod-momet approximatio i reliability of structural systems: Critical review ad alterative approach Structural Safety 983;(): 3 [0] lemaire M Fiabilité des structures Hermes 005 [] Hasofer AM, Lid NC A exact ad ivariat first order reliability format Joural of Egieerig Mechaics 974;00: [] Der Kiureghia A, Liu PL Multivariate distributio models with prescribed margials ad covariaces Probabilistic Egieerig Mechaics 986;(): 05
A probabilistic proof of a binomial identity
A probabilistic proof of a biomial idetity Joatho Peterso Abstract We give a elemetary probabilistic proof of a biomial idetity. The proof is obtaied by computig the probability of a certai evet i two
More informationI. Chi-squared Distributions
1 M 358K Supplemet to Chapter 23: CHI-SQUARED DISTRIBUTIONS, T-DISTRIBUTIONS, AND DEGREES OF FREEDOM To uderstad t-distributios, we first eed to look at aother family of distributios, the chi-squared distributios.
More informationChapter 6: Variance, the law of large numbers and the Monte-Carlo method
Chapter 6: Variace, the law of large umbers ad the Mote-Carlo method Expected value, variace, ad Chebyshev iequality. If X is a radom variable recall that the expected value of X, E[X] is the average value
More informationAsymptotic Growth of Functions
CMPS Itroductio to Aalysis of Algorithms Fall 3 Asymptotic Growth of Fuctios We itroduce several types of asymptotic otatio which are used to compare the performace ad efficiecy of algorithms As we ll
More informationProperties of MLE: consistency, asymptotic normality. Fisher information.
Lecture 3 Properties of MLE: cosistecy, asymptotic ormality. Fisher iformatio. I this sectio we will try to uderstad why MLEs are good. Let us recall two facts from probability that we be used ofte throughout
More informationOutput Analysis (2, Chapters 10 &11 Law)
B. Maddah ENMG 6 Simulatio 05/0/07 Output Aalysis (, Chapters 10 &11 Law) Comparig alterative system cofiguratio Sice the output of a simulatio is radom, the comparig differet systems via simulatio should
More informationChapter 7 Methods of Finding Estimators
Chapter 7 for BST 695: Special Topics i Statistical Theory. Kui Zhag, 011 Chapter 7 Methods of Fidig Estimators Sectio 7.1 Itroductio Defiitio 7.1.1 A poit estimator is ay fuctio W( X) W( X1, X,, X ) of
More informationDepartment of Computer Science, University of Otago
Departmet of Computer Sciece, Uiversity of Otago Techical Report OUCS-2006-09 Permutatios Cotaiig May Patters Authors: M.H. Albert Departmet of Computer Sciece, Uiversity of Otago Micah Colema, Rya Fly
More informationMaximum Likelihood Estimators.
Lecture 2 Maximum Likelihood Estimators. Matlab example. As a motivatio, let us look at oe Matlab example. Let us geerate a radom sample of size 00 from beta distributio Beta(5, 2). We will lear the defiitio
More informationTaking DCOP to the Real World: Efficient Complete Solutions for Distributed Multi-Event Scheduling
Taig DCOP to the Real World: Efficiet Complete Solutios for Distributed Multi-Evet Schedulig Rajiv T. Maheswara, Milid Tambe, Emma Bowrig, Joatha P. Pearce, ad Pradeep araatham Uiversity of Souther Califoria
More informationPROCEEDINGS OF THE YEREVAN STATE UNIVERSITY AN ALTERNATIVE MODEL FOR BONUS-MALUS SYSTEM
PROCEEDINGS OF THE YEREVAN STATE UNIVERSITY Physical ad Mathematical Scieces 2015, 1, p. 15 19 M a t h e m a t i c s AN ALTERNATIVE MODEL FOR BONUS-MALUS SYSTEM A. G. GULYAN Chair of Actuarial Mathematics
More informationModified Line Search Method for Global Optimization
Modified Lie Search Method for Global Optimizatio Cria Grosa ad Ajith Abraham Ceter of Excellece for Quatifiable Quality of Service Norwegia Uiversity of Sciece ad Techology Trodheim, Norway {cria, ajith}@q2s.tu.o
More informationConvexity, Inequalities, and Norms
Covexity, Iequalities, ad Norms Covex Fuctios You are probably familiar with the otio of cocavity of fuctios. Give a twicedifferetiable fuctio ϕ: R R, We say that ϕ is covex (or cocave up) if ϕ (x) 0 for
More informationUniversity of California, Los Angeles Department of Statistics. Distributions related to the normal distribution
Uiversity of Califoria, Los Ageles Departmet of Statistics Statistics 100B Istructor: Nicolas Christou Three importat distributios: Distributios related to the ormal distributio Chi-square (χ ) distributio.
More informationA Recursive Formula for Moments of a Binomial Distribution
A Recursive Formula for Momets of a Biomial Distributio Árpád Béyi beyi@mathumassedu, Uiversity of Massachusetts, Amherst, MA 01003 ad Saverio M Maago smmaago@psavymil Naval Postgraduate School, Moterey,
More informationCME 302: NUMERICAL LINEAR ALGEBRA FALL 2005/06 LECTURE 8
CME 30: NUMERICAL LINEAR ALGEBRA FALL 005/06 LECTURE 8 GENE H GOLUB 1 Positive Defiite Matrices A matrix A is positive defiite if x Ax > 0 for all ozero x A positive defiite matrix has real ad positive
More informationLesson 15 ANOVA (analysis of variance)
Outlie Variability -betwee group variability -withi group variability -total variability -F-ratio Computatio -sums of squares (betwee/withi/total -degrees of freedom (betwee/withi/total -mea square (betwee/withi
More informationSoving Recurrence Relations
Sovig Recurrece Relatios Part 1. Homogeeous liear 2d degree relatios with costat coefficiets. Cosider the recurrece relatio ( ) T () + at ( 1) + bt ( 2) = 0 This is called a homogeeous liear 2d degree
More information0.7 0.6 0.2 0 0 96 96.5 97 97.5 98 98.5 99 99.5 100 100.5 96.5 97 97.5 98 98.5 99 99.5 100 100.5
Sectio 13 Kolmogorov-Smirov test. Suppose that we have a i.i.d. sample X 1,..., X with some ukow distributio P ad we would like to test the hypothesis that P is equal to a particular distributio P 0, i.e.
More informationNon-life insurance mathematics. Nils F. Haavardsson, University of Oslo and DNB Skadeforsikring
No-life isurace mathematics Nils F. Haavardsso, Uiversity of Oslo ad DNB Skadeforsikrig Mai issues so far Why does isurace work? How is risk premium defied ad why is it importat? How ca claim frequecy
More informationPermutations, the Parity Theorem, and Determinants
1 Permutatios, the Parity Theorem, ad Determiats Joh A. Guber Departmet of Electrical ad Computer Egieerig Uiversity of Wiscosi Madiso Cotets 1 What is a Permutatio 1 2 Cycles 2 2.1 Traspositios 4 3 Orbits
More informationLecture 13. Lecturer: Jonathan Kelner Scribe: Jonathan Pines (2009)
18.409 A Algorithmist s Toolkit October 27, 2009 Lecture 13 Lecturer: Joatha Keler Scribe: Joatha Pies (2009) 1 Outlie Last time, we proved the Bru-Mikowski iequality for boxes. Today we ll go over the
More informationThe analysis of the Cournot oligopoly model considering the subjective motive in the strategy selection
The aalysis of the Courot oligopoly model cosiderig the subjective motive i the strategy selectio Shigehito Furuyama Teruhisa Nakai Departmet of Systems Maagemet Egieerig Faculty of Egieerig Kasai Uiversity
More information, a Wishart distribution with n -1 degrees of freedom and scale matrix.
UMEÅ UNIVERSITET Matematisk-statistiska istitutioe Multivariat dataaalys D MSTD79 PA TENTAMEN 004-0-9 LÖSNINGSFÖRSLAG TILL TENTAMEN I MATEMATISK STATISTIK Multivariat dataaalys D, 5 poäg.. Assume that
More informationChapter 5: Inner Product Spaces
Chapter 5: Ier Product Spaces Chapter 5: Ier Product Spaces SECION A Itroductio to Ier Product Spaces By the ed of this sectio you will be able to uderstad what is meat by a ier product space give examples
More informationSolutions to Selected Problems In: Pattern Classification by Duda, Hart, Stork
Solutios to Selected Problems I: Patter Classificatio by Duda, Hart, Stork Joh L. Weatherwax February 4, 008 Problem Solutios Chapter Bayesia Decisio Theory Problem radomized rules Part a: Let Rx be the
More informationOverview of some probability distributions.
Lecture Overview of some probability distributios. I this lecture we will review several commo distributios that will be used ofte throughtout the class. Each distributio is usually described by its probability
More informationLesson 17 Pearson s Correlation Coefficient
Outlie Measures of Relatioships Pearso s Correlatio Coefficiet (r) -types of data -scatter plots -measure of directio -measure of stregth Computatio -covariatio of X ad Y -uique variatio i X ad Y -measurig
More informationTHE REGRESSION MODEL IN MATRIX FORM. For simple linear regression, meaning one predictor, the model is. for i = 1, 2, 3,, n
We will cosider the liear regressio model i matrix form. For simple liear regressio, meaig oe predictor, the model is i = + x i + ε i for i =,,,, This model icludes the assumptio that the ε i s are a sample
More informationAn Efficient Polynomial Approximation of the Normal Distribution Function & Its Inverse Function
A Efficiet Polyomial Approximatio of the Normal Distributio Fuctio & Its Iverse Fuctio Wisto A. Richards, 1 Robi Atoie, * 1 Asho Sahai, ad 3 M. Raghuadh Acharya 1 Departmet of Mathematics & Computer Sciece;
More informationHere are a couple of warnings to my students who may be here to get a copy of what happened on a day that you missed.
This documet was writte ad copyrighted by Paul Dawkis. Use of this documet ad its olie versio is govered by the Terms ad Coditios of Use located at http://tutorial.math.lamar.edu/terms.asp. The olie versio
More informationCase Study. Normal and t Distributions. Density Plot. Normal Distributions
Case Study Normal ad t Distributios Bret Halo ad Bret Larget Departmet of Statistics Uiversity of Wiscosi Madiso October 11 13, 2011 Case Study Body temperature varies withi idividuals over time (it ca
More informationChapter 7 - Sampling Distributions. 1 Introduction. What is statistics? It consist of three major areas:
Chapter 7 - Samplig Distributios 1 Itroductio What is statistics? It cosist of three major areas: Data Collectio: samplig plas ad experimetal desigs Descriptive Statistics: umerical ad graphical summaries
More informationAnnuities Under Random Rates of Interest II By Abraham Zaks. Technion I.I.T. Haifa ISRAEL and Haifa University Haifa ISRAEL.
Auities Uder Radom Rates of Iterest II By Abraham Zas Techio I.I.T. Haifa ISRAEL ad Haifa Uiversity Haifa ISRAEL Departmet of Mathematics, Techio - Israel Istitute of Techology, 3000, Haifa, Israel I memory
More informationDefinition. A variable X that takes on values X 1, X 2, X 3,...X k with respective frequencies f 1, f 2, f 3,...f k has mean
1 Social Studies 201 October 13, 2004 Note: The examples i these otes may be differet tha used i class. However, the examples are similar ad the methods used are idetical to what was preseted i class.
More informationIn nite Sequences. Dr. Philippe B. Laval Kennesaw State University. October 9, 2008
I ite Sequeces Dr. Philippe B. Laval Keesaw State Uiversity October 9, 2008 Abstract This had out is a itroductio to i ite sequeces. mai de itios ad presets some elemetary results. It gives the I ite Sequeces
More informationTHE HEIGHT OF q-binary SEARCH TREES
THE HEIGHT OF q-binary SEARCH TREES MICHAEL DRMOTA AND HELMUT PRODINGER Abstract. q biary search trees are obtaied from words, equipped with the geometric distributio istead of permutatios. The average
More informationEntropy of bi-capacities
Etropy of bi-capacities Iva Kojadiovic LINA CNRS FRE 2729 Site école polytechique de l uiv. de Nates Rue Christia Pauc 44306 Nates, Frace iva.kojadiovic@uiv-ates.fr Jea-Luc Marichal Applied Mathematics
More informationFOUNDATIONS OF MATHEMATICS AND PRE-CALCULUS GRADE 10
FOUNDATIONS OF MATHEMATICS AND PRE-CALCULUS GRADE 10 [C] Commuicatio Measuremet A1. Solve problems that ivolve liear measuremet, usig: SI ad imperial uits of measure estimatio strategies measuremet strategies.
More information5 Boolean Decision Trees (February 11)
5 Boolea Decisio Trees (February 11) 5.1 Graph Coectivity Suppose we are give a udirected graph G, represeted as a boolea adjacecy matrix = (a ij ), where a ij = 1 if ad oly if vertices i ad j are coected
More informationLecture 5: Span, linear independence, bases, and dimension
Lecture 5: Spa, liear idepedece, bases, ad dimesio Travis Schedler Thurs, Sep 23, 2010 (versio: 9/21 9:55 PM) 1 Motivatio Motivatio To uderstad what it meas that R has dimesio oe, R 2 dimesio 2, etc.;
More informationIncremental calculation of weighted mean and variance
Icremetal calculatio of weighted mea ad variace Toy Fich faf@cam.ac.uk dot@dotat.at Uiversity of Cambridge Computig Service February 009 Abstract I these otes I eplai how to derive formulae for umerically
More informationUC Berkeley Department of Electrical Engineering and Computer Science. EE 126: Probablity and Random Processes. Solutions 9 Spring 2006
Exam format UC Bereley Departmet of Electrical Egieerig ad Computer Sciece EE 6: Probablity ad Radom Processes Solutios 9 Sprig 006 The secod midterm will be held o Wedesday May 7; CHECK the fial exam
More informationCooley-Tukey. Tukey FFT Algorithms. FFT Algorithms. Cooley
Cooley Cooley-Tuey Tuey FFT Algorithms FFT Algorithms Cosider a legth- sequece x[ with a -poit DFT X[ where Represet the idices ad as +, +, Cooley Cooley-Tuey Tuey FFT Algorithms FFT Algorithms Usig these
More informationEkkehart Schlicht: Economic Surplus and Derived Demand
Ekkehart Schlicht: Ecoomic Surplus ad Derived Demad Muich Discussio Paper No. 2006-17 Departmet of Ecoomics Uiversity of Muich Volkswirtschaftliche Fakultät Ludwig-Maximilias-Uiversität Müche Olie at http://epub.ub.ui-mueche.de/940/
More informationA Combined Continuous/Binary Genetic Algorithm for Microstrip Antenna Design
A Combied Cotiuous/Biary Geetic Algorithm for Microstrip Atea Desig Rady L. Haupt The Pesylvaia State Uiversity Applied Research Laboratory P. O. Box 30 State College, PA 16804-0030 haupt@ieee.org Abstract:
More information*The most important feature of MRP as compared with ordinary inventory control analysis is its time phasing feature.
Itegrated Productio ad Ivetory Cotrol System MRP ad MRP II Framework of Maufacturig System Ivetory cotrol, productio schedulig, capacity plaig ad fiacial ad busiess decisios i a productio system are iterrelated.
More informationSequences and Series
CHAPTER 9 Sequeces ad Series 9.. Covergece: Defiitio ad Examples Sequeces The purpose of this chapter is to itroduce a particular way of geeratig algorithms for fidig the values of fuctios defied by their
More informationFIBONACCI NUMBERS: AN APPLICATION OF LINEAR ALGEBRA. 1. Powers of a matrix
FIBONACCI NUMBERS: AN APPLICATION OF LINEAR ALGEBRA. Powers of a matrix We begi with a propositio which illustrates the usefuless of the diagoalizatio. Recall that a square matrix A is diogaalizable if
More informationClass Meeting # 16: The Fourier Transform on R n
MATH 18.152 COUSE NOTES - CLASS MEETING # 16 18.152 Itroductio to PDEs, Fall 2011 Professor: Jared Speck Class Meetig # 16: The Fourier Trasform o 1. Itroductio to the Fourier Trasform Earlier i the course,
More informationConfidence Intervals for One Mean
Chapter 420 Cofidece Itervals for Oe Mea Itroductio This routie calculates the sample size ecessary to achieve a specified distace from the mea to the cofidece limit(s) at a stated cofidece level for a
More informationNormal Distribution.
Normal Distributio www.icrf.l Normal distributio I probability theory, the ormal or Gaussia distributio, is a cotiuous probability distributio that is ofte used as a first approimatio to describe realvalued
More informationOverview on S-Box Design Principles
Overview o S-Box Desig Priciples Debdeep Mukhopadhyay Assistat Professor Departmet of Computer Sciece ad Egieerig Idia Istitute of Techology Kharagpur INDIA -721302 What is a S-Box? S-Boxes are Boolea
More informationStatistical inference: example 1. Inferential Statistics
Statistical iferece: example 1 Iferetial Statistics POPULATION SAMPLE A clothig store chai regularly buys from a supplier large quatities of a certai piece of clothig. Each item ca be classified either
More informationVladimir N. Burkov, Dmitri A. Novikov MODELS AND METHODS OF MULTIPROJECTS MANAGEMENT
Keywords: project maagemet, resource allocatio, etwork plaig Vladimir N Burkov, Dmitri A Novikov MODELS AND METHODS OF MULTIPROJECTS MANAGEMENT The paper deals with the problems of resource allocatio betwee
More informationTO: Users of the ACTEX Review Seminar on DVD for SOA Exam MLC
TO: Users of the ACTEX Review Semiar o DVD for SOA Eam MLC FROM: Richard L. (Dick) Lodo, FSA Dear Studets, Thak you for purchasig the DVD recordig of the ACTEX Review Semiar for SOA Eam M, Life Cotigecies
More informationMeasures of Spread and Boxplots Discrete Math, Section 9.4
Measures of Spread ad Boxplots Discrete Math, Sectio 9.4 We start with a example: Example 1: Comparig Mea ad Media Compute the mea ad media of each data set: S 1 = {4, 6, 8, 10, 1, 14, 16} S = {4, 7, 9,
More informationInstallment Joint Life Insurance Actuarial Models with the Stochastic Interest Rate
Iteratioal Coferece o Maagemet Sciece ad Maagemet Iovatio (MSMI 4) Istallmet Joit Life Isurace ctuarial Models with the Stochastic Iterest Rate Nia-Nia JI a,*, Yue LI, Dog-Hui WNG College of Sciece, Harbi
More informationThe Stable Marriage Problem
The Stable Marriage Problem William Hut Lae Departmet of Computer Sciece ad Electrical Egieerig, West Virgiia Uiversity, Morgatow, WV William.Hut@mail.wvu.edu 1 Itroductio Imagie you are a matchmaker,
More informationExample 2 Find the square root of 0. The only square root of 0 is 0 (since 0 is not positive or negative, so those choices don t exist here).
BEGINNING ALGEBRA Roots ad Radicals (revised summer, 00 Olso) Packet to Supplemet the Curret Textbook - Part Review of Square Roots & Irratioals (This portio ca be ay time before Part ad should mostly
More informationNATIONAL SENIOR CERTIFICATE GRADE 12
NATIONAL SENIOR CERTIFICATE GRADE MATHEMATICS P EXEMPLAR 04 MARKS: 50 TIME: 3 hours This questio paper cosists of 8 pages ad iformatio sheet. Please tur over Mathematics/P DBE/04 NSC Grade Eemplar INSTRUCTIONS
More informationAnalyzing Longitudinal Data from Complex Surveys Using SUDAAN
Aalyzig Logitudial Data from Complex Surveys Usig SUDAAN Darryl Creel Statistics ad Epidemiology, RTI Iteratioal, 312 Trotter Farm Drive, Rockville, MD, 20850 Abstract SUDAAN: Software for the Statistical
More informationPlug-in martingales for testing exchangeability on-line
Plug-i martigales for testig exchageability o-lie Valetia Fedorova, Alex Gammerma, Ilia Nouretdiov, ad Vladimir Vovk Computer Learig Research Cetre Royal Holloway, Uiversity of Lodo, UK {valetia,ilia,alex,vovk}@cs.rhul.ac.uk
More informationResearch Article Sign Data Derivative Recovery
Iteratioal Scholarly Research Network ISRN Applied Mathematics Volume 0, Article ID 63070, 7 pages doi:0.540/0/63070 Research Article Sig Data Derivative Recovery L. M. Housto, G. A. Glass, ad A. D. Dymikov
More informationLecture 3. denote the orthogonal complement of S k. Then. 1 x S k. n. 2 x T Ax = ( ) λ x. with x = 1, we have. i = λ k x 2 = λ k.
18.409 A Algorithmist s Toolkit September 17, 009 Lecture 3 Lecturer: Joatha Keler Scribe: Adre Wibisoo 1 Outlie Today s lecture covers three mai parts: Courat-Fischer formula ad Rayleigh quotiets The
More informationCOMPARISON OF THE EFFICIENCY OF S-CONTROL CHART AND EWMA-S 2 CONTROL CHART FOR THE CHANGES IN A PROCESS
COMPARISON OF THE EFFICIENCY OF S-CONTROL CHART AND EWMA-S CONTROL CHART FOR THE CHANGES IN A PROCESS Supraee Lisawadi Departmet of Mathematics ad Statistics, Faculty of Sciece ad Techoology, Thammasat
More information1 Computing the Standard Deviation of Sample Means
Computig the Stadard Deviatio of Sample Meas Quality cotrol charts are based o sample meas ot o idividual values withi a sample. A sample is a group of items, which are cosidered all together for our aalysis.
More informationCS103X: Discrete Structures Homework 4 Solutions
CS103X: Discrete Structures Homewor 4 Solutios Due February 22, 2008 Exercise 1 10 poits. Silico Valley questios: a How may possible six-figure salaries i whole dollar amouts are there that cotai at least
More informationDiscrete Mathematics and Probability Theory Spring 2014 Anant Sahai Note 13
EECS 70 Discrete Mathematics ad Probability Theory Sprig 2014 Aat Sahai Note 13 Itroductio At this poit, we have see eough examples that it is worth just takig stock of our model of probability ad may
More informationCHAPTER 7: Central Limit Theorem: CLT for Averages (Means)
CHAPTER 7: Cetral Limit Theorem: CLT for Averages (Meas) X = the umber obtaied whe rollig oe six sided die oce. If we roll a six sided die oce, the mea of the probability distributio is X P(X = x) Simulatio:
More informationHypothesis testing. Null and alternative hypotheses
Hypothesis testig Aother importat use of samplig distributios is to test hypotheses about populatio parameters, e.g. mea, proportio, regressio coefficiets, etc. For example, it is possible to stipulate
More informationwhere: T = number of years of cash flow in investment's life n = the year in which the cash flow X n i = IRR = the internal rate of return
EVALUATING ALTERNATIVE CAPITAL INVESTMENT PROGRAMS By Ke D. Duft, Extesio Ecoomist I the March 98 issue of this publicatio we reviewed the procedure by which a capital ivestmet project was assessed. The
More informationWeek 3 Conditional probabilities, Bayes formula, WEEK 3 page 1 Expected value of a random variable
Week 3 Coditioal probabilities, Bayes formula, WEEK 3 page 1 Expected value of a radom variable We recall our discussio of 5 card poker hads. Example 13 : a) What is the probability of evet A that a 5
More informationDecomposition of Gini and the generalized entropy inequality measures. Abstract
Decompositio of Gii ad the geeralized etropy iequality measures Stéphae Mussard LAMETA Uiversity of Motpellier I Fraçoise Seyte LAMETA Uiversity of Motpellier I Michel Terraza LAMETA Uiversity of Motpellier
More information5: Introduction to Estimation
5: Itroductio to Estimatio Cotets Acroyms ad symbols... 1 Statistical iferece... Estimatig µ with cofidece... 3 Samplig distributio of the mea... 3 Cofidece Iterval for μ whe σ is kow before had... 4 Sample
More informationDAME - Microsoft Excel add-in for solving multicriteria decision problems with scenarios Radomir Perzina 1, Jaroslav Ramik 2
Itroductio DAME - Microsoft Excel add-i for solvig multicriteria decisio problems with scearios Radomir Perzia, Jaroslav Ramik 2 Abstract. The mai goal of every ecoomic aget is to make a good decisio,
More information1. MATHEMATICAL INDUCTION
1. MATHEMATICAL INDUCTION EXAMPLE 1: Prove that for ay iteger 1. Proof: 1 + 2 + 3 +... + ( + 1 2 (1.1 STEP 1: For 1 (1.1 is true, sice 1 1(1 + 1. 2 STEP 2: Suppose (1.1 is true for some k 1, that is 1
More informationSECTION 1.5 : SUMMATION NOTATION + WORK WITH SEQUENCES
SECTION 1.5 : SUMMATION NOTATION + WORK WITH SEQUENCES Read Sectio 1.5 (pages 5 9) Overview I Sectio 1.5 we lear to work with summatio otatio ad formulas. We will also itroduce a brief overview of sequeces,
More informationTHIN SEQUENCES AND THE GRAM MATRIX PAMELA GORKIN, JOHN E. MCCARTHY, SANDRA POTT, AND BRETT D. WICK
THIN SEQUENCES AND THE GRAM MATRIX PAMELA GORKIN, JOHN E MCCARTHY, SANDRA POTT, AND BRETT D WICK Abstract We provide a ew proof of Volberg s Theorem characterizig thi iterpolatig sequeces as those for
More information2 MATH 101B: ALGEBRA II, PART D: REPRESENTATIONS OF GROUPS
2 MATH 101B: ALGEBRA II, PART D: REPRESENTATIONS OF GROUPS 1. The group rig k[g] The mai idea is that represetatios of a group G over a field k are the same as modules over the group rig k[g]. First I
More informationTHE ARITHMETIC OF INTEGERS. - multiplication, exponentiation, division, addition, and subtraction
THE ARITHMETIC OF INTEGERS - multiplicatio, expoetiatio, divisio, additio, ad subtractio What to do ad what ot to do. THE INTEGERS Recall that a iteger is oe of the whole umbers, which may be either positive,
More informationPresent Values, Investment Returns and Discount Rates
Preset Values, Ivestmet Returs ad Discout Rates Dimitry Midli, ASA, MAAA, PhD Presidet CDI Advisors LLC dmidli@cdiadvisors.com May 2, 203 Copyright 20, CDI Advisors LLC The cocept of preset value lies
More informationIrreducible polynomials with consecutive zero coefficients
Irreducible polyomials with cosecutive zero coefficiets Theodoulos Garefalakis Departmet of Mathematics, Uiversity of Crete, 71409 Heraklio, Greece Abstract Let q be a prime power. We cosider the problem
More informationRanking Irregularities When Evaluating Alternatives by Using Some ELECTRE Methods
Please use the followig referece regardig this paper: Wag, X., ad E. Triataphyllou, Rakig Irregularities Whe Evaluatig Alteratives by Usig Some ELECTRE Methods, Omega, Vol. 36, No. 1, pp. 45-63, February
More informationLECTURE 13: Cross-validation
LECTURE 3: Cross-validatio Resampli methods Cross Validatio Bootstrap Bias ad variace estimatio with the Bootstrap Three-way data partitioi Itroductio to Patter Aalysis Ricardo Gutierrez-Osua Texas A&M
More informationSAMPLE QUESTIONS FOR FINAL EXAM. (1) (2) (3) (4) Find the following using the definition of the Riemann integral: (2x + 1)dx
SAMPLE QUESTIONS FOR FINAL EXAM REAL ANALYSIS I FALL 006 3 4 Fid the followig usig the defiitio of the Riema itegral: a 0 x + dx 3 Cosider the partitio P x 0 3, x 3 +, x 3 +,......, x 3 3 + 3 of the iterval
More informationInference on Proportion. Chapter 8 Tests of Statistical Hypotheses. Sampling Distribution of Sample Proportion. Confidence Interval
Chapter 8 Tests of Statistical Hypotheses 8. Tests about Proportios HT - Iferece o Proportio Parameter: Populatio Proportio p (or π) (Percetage of people has o health isurace) x Statistic: Sample Proportio
More information1 The Gaussian channel
ECE 77 Lecture 0 The Gaussia chael Objective: I this lecture we will lear about commuicatio over a chael of practical iterest, i which the trasmitted sigal is subjected to additive white Gaussia oise.
More informationThe following example will help us understand The Sampling Distribution of the Mean. C1 C2 C3 C4 C5 50 miles 84 miles 38 miles 120 miles 48 miles
The followig eample will help us uderstad The Samplig Distributio of the Mea Review: The populatio is the etire collectio of all idividuals or objects of iterest The sample is the portio of the populatio
More informationThe Gompertz Makeham coupling as a Dynamic Life Table. Abraham Zaks. Technion I.I.T. Haifa ISRAEL. Abstract
The Gompertz Makeham couplig as a Dyamic Life Table By Abraham Zaks Techio I.I.T. Haifa ISRAEL Departmet of Mathematics, Techio - Israel Istitute of Techology, 32000, Haifa, Israel Abstract A very famous
More informationYour organization has a Class B IP address of 166.144.0.0 Before you implement subnetting, the Network ID and Host ID are divided as follows:
Subettig Subettig is used to subdivide a sigle class of etwork i to multiple smaller etworks. Example: Your orgaizatio has a Class B IP address of 166.144.0.0 Before you implemet subettig, the Network
More informationTheorems About Power Series
Physics 6A Witer 20 Theorems About Power Series Cosider a power series, f(x) = a x, () where the a are real coefficiets ad x is a real variable. There exists a real o-egative umber R, called the radius
More informationNow here is the important step
LINEST i Excel The Excel spreadsheet fuctio "liest" is a complete liear least squares curve fittig routie that produces ucertaity estimates for the fit values. There are two ways to access the "liest"
More informationA Faster Clause-Shortening Algorithm for SAT with No Restriction on Clause Length
Joural o Satisfiability, Boolea Modelig ad Computatio 1 2005) 49-60 A Faster Clause-Shorteig Algorithm for SAT with No Restrictio o Clause Legth Evgey Datsi Alexader Wolpert Departmet of Computer Sciece
More informationPSYCHOLOGICAL STATISTICS
UNIVERSITY OF CALICUT SCHOOL OF DISTANCE EDUCATION B Sc. Cousellig Psychology (0 Adm.) IV SEMESTER COMPLEMENTARY COURSE PSYCHOLOGICAL STATISTICS QUESTION BANK. Iferetial statistics is the brach of statistics
More informationA Mathematical Perspective on Gambling
A Mathematical Perspective o Gamblig Molly Maxwell Abstract. This paper presets some basic topics i probability ad statistics, icludig sample spaces, probabilistic evets, expectatios, the biomial ad ormal
More informationGCSE STATISTICS. 4) How to calculate the range: The difference between the biggest number and the smallest number.
GCSE STATISTICS You should kow: 1) How to draw a frequecy diagram: e.g. NUMBER TALLY FREQUENCY 1 3 5 ) How to draw a bar chart, a pictogram, ad a pie chart. 3) How to use averages: a) Mea - add up all
More informationChapter 7: Confidence Interval and Sample Size
Chapter 7: Cofidece Iterval ad Sample Size Learig Objectives Upo successful completio of Chapter 7, you will be able to: Fid the cofidece iterval for the mea, proportio, ad variace. Determie the miimum
More informationBASIC STATISTICS. f(x 1,x 2,..., x n )=f(x 1 )f(x 2 ) f(x n )= f(x i ) (1)
BASIC STATISTICS. SAMPLES, RANDOM SAMPLING AND SAMPLE STATISTICS.. Radom Sample. The radom variables X,X 2,..., X are called a radom sample of size from the populatio f(x if X,X 2,..., X are mutually idepedet
More informationFunction factorization using warped Gaussian processes
Fuctio factorizatio usig warped Gaussia processes Mikkel N. Schmidt ms@imm.dtu.dk Uiversity of Cambridge, Departmet of Egieerig, Trumpigto Street, Cambridge, CB2 PZ, UK Abstract We itroduce a ew approach
More information