scieee AI-readable full text Open interactive document viewer

The limit relations between generalized orthogonal polynomials

Álvarez Nodarse, Renato; Marcellán Español, Francisco

Abstract

We consider the different limit transitions for modifications of the classical polynomials obtained by the addition of one or two point masses at the ends of the interval of orthogonality. The connections between Jacobi, Laguerre, Charlier, Meixner, Kravchuk and Hahn generalized polynomials are established.

Full text

LIMIT RELATIONS BETWEEN GENERALIZED ORTHOGONAL POLYNOMIALS . 1 R.  Alvarez-No darse 2 and F. Marcellan 3 Departamento de Matematicas. Escuela Politecnica Superior. Universidad Carlos III de Madrid. Butarque 15, 28911,Leganes,Madrid. Key words and phrases: Classical Orthogonal p olynomials, Limit relations. AMS (MOS) sub ject classication: 33A65 Abstract We consider the dierent limit transition for mo dications of the classical p olynomials via the addition of one or two p oint masses at the ends of the interval of orthogonality. The connections b etween Jacobi, Laguerre, Charlier, Meixner, Kravchuk and Hahn generalized p olynomial s are established. 1 Intro duction. Polynomials orthogonal with resp ect to measures which are more general than weight functions app ear as eigenfunctions of a fourth order linear dierential op erator with p olynomial co ecients. This sp ectral approach leads to Laguerre-typ e, Legendre-typ e and Jacobi-typ e p olynomials intro duced by H.L.Krall [20]. A general analysis when a mo dication of a linear functional in the linear space of p olynomials with real co ecients via the addition of one delta Dirac measure was started by Chihara [9] in the p ositive denite case and Marcellan and Maroni [22] for quasi-denite linear functionals. For two p oint masses there exist very few examples in the literature. (see [19], [11], [17] and [21]) A sp ecial emphasis was given to the mo dications of classical linear functionals (Hermite, Laguerre, Jacobi and Bessel) in the framework of the so-called semiclassical orthogonal p olynomials. For discrete orthogonal p olynomials, Bavinck and van Haeringen [7] obtained an innite order dierence equation for generalized Meixner p olynomials, i.e., p olynomials orthogonal with resp ect to the mo dication of the Meixner weight with a p oint mass at x = 0. The same was found for generalized Charlier p olynomials by Bavinck and Ko eko ek [8]. In a series of pap ers [2]-[4] we obtained the representation as hyp ergeometric functions for generalized Meixner, Charlier, Kravchuk and Hahn p olynomials as well as the corresp onding second order dierence equation that such p olynomials satisfy. Notice that the co ecients of those dierence equations are p olynomials of xed degree and they dep end on n as a 1 Octob er 29, 1996 2 E-mail: [email protected] 3 E-mail: [email protected] 1 parameter. The aim of the present contribution is to obtain an analogue of the Askey tableau for such a kind of generalized p olynomials with the description of the continuous generalized orthogonal p olynomials as limit case of the discrete generalized orthogonal p olynomials. Furthermore, we deduce the explicit second order linear dierential equations for two examples which attracted the interest of the researchers: the Laguerre [13] and the Jacobi [19] case. In Section 2 we present a summary of the more useful prop erties of classical p olynomials b oth in the discrete and continuous case. Section 3 is devoted to an explicit representation of generalized p olynomials in terms of the classical ones when we add one p oint mass at zero (Laguerre, Meixner, Charlier, Kravchuk) or two mass p oints at the ends of the convex hull of the supp ort of the measure (Jacobi and Hahn). Further, we obtain the explicit expression for second order dierential equations (SODE) in the cases of Laguerre and Jacobi. Notice that this SODE was found in [13] for Laguerre case while for the Jacobi case [19] the co ecients were not deduced explicitly. Moreover, an innite order equation for the Laguerre case was found in [13] as well as for the Gegenbauer case in [16]. In Section 4 the continuous case is obtained as a limit of the discrete one, as well as the dierent transitions b etween the discrete families. 2 Some Preliminary Results. In this Section we have summarized some formulas of the classical orthogonal monic p olynomials ( P n ( x ) = x n + ::: ) which we will use later on. These p olynomials are orthogonal with resp ect to a linear functional C on the linear space of p olynomials with real co ecients dened as ( IN = f 0 ; 1 ; 2 ; ::: g ) < C ; P > = 8 > > > > < > > > > : X x 2 IN  ( x ) P ( x ) ; Hahn, Meixner, Kravchuk and Charlier Z b a  ( x ) P ( x ) dx; Jacobi and Laguerre (1) where  ( x ) is a weight function satisfying a Pearson equation. In the continuous case it has the form d dx [  ( x )  ( x )] =  ( x )  ( x ) : They satisfy a second order dierential equation of hyp ergeometric typ e  ( x ) P 00 n ( x ) +  ( x ) P 0 n ( x ) +  n P n ( x ) = 0 ; (2) where  ( x ) is a p olynomial of degree 1 and  ( x ) is a p olynomial of degree at most 2, such that  ( x ) vanishes at the ends of the interval of orthogonality. The p olynomial solutions of equation (2) are uniquely determined, up to a normalized factor ( R n ), by the Ro drigues formula (see [23] page 4 Eq.(1.2.8)): 2 P n ( x ) = R n  ( x ) d n d x n [  n ( x )  ( x )] : (3) In the discrete case, the Pearson-typ e dierence equation has the form 4 [  ( x )  ( x )] =  ( x )  ( x ) ; where 5 f ( x ) = f ( x )  f ( x  1) ; 4 f ( x ) = f ( x + 1)  f ( x ) : The Pearson-typ e dierence equation can b e written in the equivalent form  ( x + 1)  ( x ) =  ( x ) +  ( x )  ( x + 1) : In this case instead of a dierential equation, they satisfy a second order dierence equation of hyp ergeometric typ e  ( x ) 4 5 P n ( x ) +  ( x ) 4 P n ( x ) +  n P n ( x ) = 0 ; (4) where  ( x ) is also a p olynomial of degree 1 and  ( x ) is a p olynomial of degree at most 2, such that  ( x ) vanishes at one of the ends of the convex hull of the supp ort and  ( x ) +  ( x ) vanishes in the other. The p olynomial solutions of equation (4) are uniquely determined, up to a normalized factor ( R n ), by the dierence analog of the Ro drigues formula (see [23] page 24 Eq.(2.2.7)): P n ( x ) = R n  ( x ) 5 n "  ( x + n ) n Y k =1  ( x + k ) # : (5) The orthogonality with resp ect to the linear funtional C means that < C ; P n P m > = 8 < : 0 m 6 = n d 2 n m = n : (6) In b oth cases, they satisfy a three term recurrence relation of the form xP n ( x ) =  n P n +1 ( x ) +  n P n ( x ) +  n P n  1 ( x ) ; n  0 P  1 ( x ) = 0 and P 0 ( x ) = 1 (7) and the Christoel-Darb oux formula n  1 X m =0 P m ( x ) P m ( y ) d 2 m = 1 x  y a n  1 a n P n ( x ) P n  1 ( y )  P n ( y ) P n  1 ( x ) d 2 n  1 n = 1 ; 2 ; 3 ; ::: : (8) Here a n is the leading co ecient of the p olynomial, i.e., the co ecient of the nth p ower of x in the expansion (in our cases since P n is monic a n = 1) P n ( x ) = a n x n + b n x n  1 + ::: = x n + b n x n  1 + ::: (9) We will consider the mo dication of the following classical monic orthogonal p olynomials. 1. The discrete case. 3 1. The Meixner p olynomials M  ; n ( x ), orthogonal with resp ect to the weight function  ( x ) supp orted on [0 ; 1 ), where  ( x ) = x;  ( x ) =    x (1   ) ; 0 <  < 1 ;  > 0 ;  n = n (1   ) ; and R n = 1 (   1) n ;  ( x ) =  x (1   )  (  + x ) (  )(1 + x ) ; d 2 n = n !(  ) n  n (1   ) 2 n : 2. The Kravchuk p olynomials K p n ( x ), orthogonal with resp ect to the weight function  ( x ) supp orted on [0 ; N ], with n  N  ( x ) = x;  ( x ) = N p  x 1  p ; 0 < p < 1 ;  n = n 1  p ; and R n = ( p  1) n ;  ( x ) = p x N !(1  p ) N  x ( N + 1  x )(1 + x ) ; d 2 n = n ! N ! p n (1  p ) n ( N  n )! : 3. The Charlier p olynomials C  n ( x ), orthogonal with resp ect to the weight function  ( x ) supp orted on [0 ; 1 ), where  ( x ) = x; ;  ( x ) =   x ;  > 0 ;  n = n; and R n = (  1) n ;  ( x ) =  x e   (1 + x ) ; d 2 n = n !  n : 4. The Hahn p olynomials h ; n ( x; N ), orthogonal with resp ect to the weight function  ( x ) supp orted on [0 ; N ), where (  >  1 ;  >  1)  ( x ) = x ( x +   N ) ;  ( x ) = (  + 1)( N  1)  x (  +  + 2) ;  n = n (  +  + N + 1) ; and R n = (  1) n (  +  + n + 1) n ;  ( x ) = ( N )(  +  + 2)(  + N  x )(  + 1 + x ) (  + 1)(  + 1)(  +  + N + 1)( N  x )(1 + x ) ; d 2 n = ( N )(  +  + 2) n !(  + n + 1)(  + n + 1)(  +  + N + n + 1)(  +  + n + 1)  2 n (  + 1)(  + 1)(  +  + N + 1)(  +  + 2 n + 1)( N  n  1)!(  +  + n + 1) : They satisfy the symmetry prop erty h  ; n ( N  1  x; N ) = (  1) n h ; n ( x; N ) : (10) 2. The continuous case. 4 1. The Jacobi p olynomials P ; n ( x ), orthogonal with resp ect to the weight function  ( x ) supp orted on [  1 ; 1], where  ( x ) = 1  x 2 ;  ( x ) =  (  +  + 2) x +    ;  n = n ( n +  +  + 1) ; R n = (  1) n ( n +  +  + 1) n ;  ( x ) = (  +  + 2) 2  +  +1 (  + 1)(  + 1) (1  x )  (1 + x )   >  1 ;  >  1 ; d 2 n = 2 2 n n !( n +  + 1)( n +  + 1)(  +  + 2) (  + 1)(  + 1)( n +  +  + 1)(2 n +  +  + 1)( n +  +  + 1) 2 n : They satisfy the symmetry prop erty P  ; n (  x ) = (  1) n P ; n ( x ) : (11) 2. The Laguerre p olynomials L  n ( x ), orthogonal with resp ect to the weight function  ( x ) supp orted on [0 ; 1 ), where  ( x ) = x;  ( x ) =  x +  + 1 ;  n = n; and R n = (  1) n ;  ( x ) = x  e  x (  + 1)  >  1 ; d 2 n = ( n +  + 1) n ! (  + 1) : In the ab ove formulas we have scaled the weight functions  ( x ) such that they b ecomes probability measures, i.e., total weight equal 1. This will b e useful in order to obtain the right limits b etween the corresp onding generalized p olynomials. For all those monic p olynomials we also know the values M  ; n (0) =  n (   1) n ( n +  ) (  ) ; K p n (0) = (  p ) n N ! ( N  n )! ; C  n (0) = (   ) n ; h ; n (0 ; N ) = (  1) n (  + n + 1)( N  1)! (  + 1)( N  n  1)!( n +  +  + 1) n ; h ; n ( N  1 ; N ) = (  + n + 1)( N  1)! (  + 1)( N  n  1)!( n +  +  + 1) n ; P ; n (1) = 2 n (  + 1) n ( n +  +  + 1) n ; P ; n (  1) = 2 n (  1) n (  + 1) n ( n +  +  + 1) n ; L  n (0) = (  1) n ( n +  + 1) (  + 1) : (12) From the hyp ergeometric representation of Jacobi p olynomials (see [23] - [25]) we can obtain the following two expressions [24] P ; +1 n  1 ( x ) = (2 n +  +  )(1  x ) 2 n (  + n ) d P ; n dx ( x ) + (2 n +  +  ) 2(  + n ) P ; n ( x ) (13) 5 and P  +1 ; n  1 ( x ) = (2 n +  +  )( x + 1) 2 n (  + n ) d P ; n dx ( x )  (2 n +  +  ) 2(  + n ) P ; n ( x ) : (14) For the kernels of the Charlier, Meixner, Kravchuk, Hahn, Jacobi and Laguerre p olynomials we have the following representation (see for instance [2]-[5] and [25]) 1. Meixner case K er M n  1 ( x; 0)  n  1 X m =0 M  ; m ( x ) M  ; m (0) d 2 m = (  1) n  1 (1   ) n  1 n ! 5 M  ; n ( x ) ; (15) K er M n  1 (0 ; 0) = n  1 X m =0 (  ) m  m m ! : (16) 2. Kravchuk case K er K n  1 ( x; 0)  n  1 X m =0 K p m ( x ) K p m (0) d 2 m = ( p  1) 1  n n ! 5 K p n ( x ) ; (17) K er K n  1 (0 ; 0) = n  1 X m =0 p m N ! (1  p ) m m !( N  m )! : (18) 3. Charlier case K er C n  1 ( x; 0)  n  1 X m =0 C  m ( x ) C  m (0) d 2 m = (  1) n  1 n ! 5 C  n ( x ) ; (19) K er C n  1 (0 ; 0) = n  1 X m =0  m m ! : (20) 4. Hahn case K er H;; n  1 ( x; 0)  n  1 X m =0 h ; m ( x; N ) h ; m (0 ; N ) d 2 m =  n ( ;  ) 5 h   1 ; n ( x; N ) ; K er H;; n  1 ( x; N  1) =  n (  ;  )(  1) n +1 4 h ;  1 n ( x; N ) : (21) where  n ( ;  ) denotes the following quantity  n ( ;  ) = (  1) n  1 (  +  + 2 n )(  + 1)(  +  + N + 1) n !(  + n )(  +  + n + N )(  +  + 2) ; (22) K er H;; n  1 (0 ; 0) = n  1 X m =0 ( m +  + 1)( m +  +  + 1) m !(  + 1)( N  m  1)!   (2 m +  +  + 1)( N  1)!(  + 1)(  +  + N + 1) (  + m + 1)(  +  + N + m + 1)(  +  + 2) ; (23) 6 K er H;; n  1 (0 ; N  1) = = n  1 X m =0 (  1) m ( m +  +  + 1)(2 m +  +  + 1)( N  1)!(  +  + N + 1 m !( N  m  1)!(  +  + N + m + 1)(  +  + 2) ; (24) and, nally, from the symmetry of the Hahn p olynomials (10) we obtain K er H;; n  1 ( N  1 ; N  1) = K er H; ; n  1 (0 ; 0) : 5. Laguerre case K er L n  1 ( x; 0)  n  1 X m =0 L  m ( x ) L  m (0) d 2 m = (  1) n  1 n ! ( L  n ) 0 ( x ) ; (25) K er L n  1 (0 ; 0) = n  1 X m =0 (  + 1) m m ! = (  + 2) n  1 ( n  1)! : (26) 6. Jacobi case K er J;; n  1 ( x;  1)  n  1 X m =0 P ; m ( x ) P ; m (  1) d 2 m =  ; n d dx P   1 ; n ( x ) ; (27) K er J;; n  1 ( x; 1)  n  1 X m =0 P ; m ( x ) P ; m (1) d 2 m = (  1) n +1   ; n d dx P ;  1 n ( x ) ; (28) where  ; n ;   ; n denote the quantities  ; n = (  1) n  1 (2 n +  +  )(  + 1) 2 n  1 n !(  + n )(  + 1)(  +  + 2) ;   ; n = (  1) n  1 (2 n +  +  )(  + 1) 2 n  1 n !(  + n )(  + 1)(  +  + 2) : (29) K er J;; n  1 (  1 ;  1) = n  1 X m =0 (  + m + 1)(  +  + m + 1)(2 m +  +  + 1)(  + 1) 2 n  1 m !(  + 1)(  + m + 1)(  +  + 2) = = (  + n + 1)(  +  + n + 1)(  + 1) 2 n  1 ( n  1)!(  + 2)(  + n )(  +  + 2) ; (30) and K er J;; n  1 (  1 ; 1) = n  1 X m =0 (  1) m (  +  + m + 1)(2 m +  +  + 1) 2 n  1 m !(  +  + 2) = = (  1) n  1 (  +  + n + 1) 2 n  1 ( n  1)! : (31) 7 and, nally, from the symmetry prop erty of the Jacobi (11) p olynomials we have K er J;; n  1 (1 ; 1) = K er J; ; n  1 (  1 ;  1) : Using the relations (13)-(14) we also obtain the following equivalent formulas for the kernels (27) and (28) K er J;; n  1 ( x;  1) = ~  ; n  (1  x ) dP ; n ( x ) dx + n P ; n ( x )  ; (32) K er J;; n  1 ( x; 1) = (  1) n +1 ~   ; n  (1 + x ) dP ; n ( x ) dx  n P ; n ( x )  ; (33) where ~  ; n ; ~   ; n denotes the quantities ~  ; n =  (  1) n (2 n +  +  + 1)(  + 1) 2 n n !(  + n + 1)(  +  + 2) ; ~   ; n =  (  1) n (2 n +  +  + 1)(  + 1) 2 n n !(  + n + 1)(  +  + 2) : (34) 3 The denition and the representation. Firstly, we will consider the case when we add a p oint mass at x = 0. This case corresp onds to the Laguerre, Charlier, Meixner and Kravchuk p olynomials. Later on, we will consider the Jacobi and Hahn p olynomials which involve two p oint masses at the ends of the interval of orthogonality. The reason of such a choice of the p oint in which we will add our p ositive mass will b e clear from formulas (39) and (41) from b elow, b ecause in such formulas app ears the value of the kernel p olynomials K n ( x; y ) and they have a very simple analytical expression in the case when y takes the values of the zeros of  ( x ) (for the continuous case) or one of the zeros of  ( x ) and  ( x ) +  ( x ) (for the discrete case). In fact this gives us a simple expression for the kernels in terms of the same p olynomials, its derivatives or dierence-derivatives (see (15)-(31)). 3.1 The Case of one p oint mass. Consider the linear functional U on the linear space of p olynomials with real co ecients dened as < U ; P > = < C ; P > + AP (0) ;;A  0 ; (35) where C is a classical moment functional (1) asso ciated to some Meixner, Charlier and Kravchuk p olynomials of a discrete variable and Laguerre p olynomials, resp ectively. We will determine the monic p olynomials P A n ( x ) which are orthogonal with resp ect to the functional U and we will prove that they exist for all p ositive A (see (40) from b elow). To obtain this, we can write the Fourier expansion of such generalized p olynomials P A n ( x ) = P n ( x ) + n  1 X k =0 a n;k P k ( x ) ; (36) 8 where P n denotes the classical monic orthogonal p olynomial (CMOP) of degree n . In order to nd the unknown co ecients a n;k we will use the orthogonality of the p olynomials P A n ( x ) with resp ect to U , i.e., < U ; P A n ( x ) P k ( x ) > = 0 8 k < n: Now putting (36) in (35) we nd: < U ; P A n ( x ) P k ( x ) > = < C ; P A n ( x ) P k ( x ) > + AP A n (0) P k (0) : (37) If we use the decomp osition (36) and taking into account the orthogonality of the classical orthogonal p olynomials with resp ect to the linear functional C , then the co ecients a n;k are given by a n;k =  A P A n (0) P k (0) d 2 k : (38) Finally the equation (36) provides us the expression P A n ( x ) = P n ( x )  AP A n (0) n  1 X k =0 P k (0) P k ( x ) d 2 k = P n ( x )  AP A n (0) K er n  1 ( x; 0) : (39) From (39) we can conclude that the representation of P A n ( x ) exists for any p ositive value of the mass A . To obtain this it is enough to evaluate (39) in x = 0, 1 + A n  1 X k =0 ( P k (0)) 2 d 2 k ! P A n (0) = P n (0) 6 = 0 ; (40) and use the fact that 1 + A n  1 X k =0 ( P k (0)) 2 d 2 k > 0 n = 1 ; 2 ; 3 ; ::: From (40) we can deduce the values of P A n (0) as follows P A n (0) = P n (0) 1 + A n  1 X k =0 ( P k (0)) 2 d 2 k : (41) From (39) and taking into account formulas (15)-(25) as well as (41), we obtain the following expressions for the generalized p olynomials (for more details see [2],[3], [5] and [13]) For Meixner p olynomials M  ;;A n ( x ) = M  ; n ( x ) + B n 5 M  ; n ( x ) = ( I + B n 5 ) M  ; n ( x ) ; (42) B n = A  n (1   )  1 (  ) n n !(1 + AK er M n  1 (0 ; 0)) : For Kravchuk p olynomials 9 Taking the limit A; B ! 0 we obtain lim A;B ! 0 ~  n ( x ) =  1  x 2  2 =  ( x ) 2 ; lim A;B ! 0 ~  n ( x ) = n (1 +  +  + n )  1  x 2  =  ( x )  n ; lim A;B ! 0 ~  n ( x ) =  1  x 2  (   +   2 x   x   x ) =  ( x )  ( x ) : 4 Limit relations b etween mo dications of orthogonal p olynomials. In this Section we will study limit relations involving the mo dications of the Jacobi and Laguerre p olynomials as well as the mo dications of the classical p olynomials of discrete variables. In some way we will obtain an analogue of the Askey-scheme of hyp ergeometric p olynomials (for a review see [18]). Results are predictible but we have found nothing of this kind in the literature. Anyway, we want to remark that the main dierence with resp ect to the classical case is the fact, as we will show b elow, that the p oint masses change. 4.1 Limit Meixner ! Laguerre. The limit relation b etween the classical Meixner and Laguerre p olynomials is well known lim h ! 0 h n M  +1 ; 1  h n  x h  = L  n ( x ) : (67) In order to obtain the analogues of this relation for generalized p olynomials we notice that (see (16) and (26)) K er M n  1 (0 ; 0) = n  1 X k =0 [ M  +1 ; 1  h k (0)] 2 d 2 k = n  1 X k =0 (  + 1) k (1  h ) k k ! : Then K er M n  1 (0 ; 0) = K er L n  1 (0 ; 0) + O ( h ). Now from the representation formulas (42) we nd M  +1 ; 1  h;A n  x h  = M  +1 ; 1  h n  x h  + A (  + 1) n (1  h ) n n ! (1 + AK er L n  1 (0 ; 0))   M  +1 ; 1  h;A n  x h   M  +1 ; 1  h;A n  x  h h  h : Multiplying this expression by the factor h n , taking the limit when h ! 0 and using (67) we notice that the right side of the last expression b ecomes into the right side of (45). Then, the following relation holds lim h ! 0 h n M  +1 ; 1  h;A n  x h  = L ;A n ( x ) : (68) 4.2 Limit Meixner ! Charlier. We start again from the classical limit relation for monic Meixner and Charlier p olynomials lim  !1 M  ;   +  n ( x ) = C  n ( x ) : (69) 16 For the kernels of Meixner p olynomials we have (see (16) and (20)) lim  !1 K er M n  1 (0 ; 0) = n  1 X k =0  k k ! = K er C n  1 (0 ; 0) : Now from formula (42) we nd that lim  !1 B n = A  n n !(1 + AK er C n  1 (0 ; 0)) ; which agrees with D n in the representation formula for Charlier p olynomials (44). Now, like in the previous case, we take the limit  ! 1 . Hence, using (69) the following relation holds lim  !1 M  ;   +  ;A n ( x ) = C ;A n ( x ) : (70) 4.3 Limit Kravchuk ! Charlier. In this case the limit relation takes the form lim N !1 K  N n ( x ) = C  n ( x ) : (71) First of all, since N ! ( N  n )!  N n then lim N !1 N n ( N  n )! N ! = 1. Using these two relations we nd that lim N !1 K er K n  1 (0 ; 0) = K er C n  1 (0 ; 0), and also from (43) we have lim N !1 A n = A  n n !(1 + AK er C n  1 (0 ; 0)) : Then from (44) we conclude that lim N !1 K  N ;A n ( x ) = C ;A n ( x ) : (72) 4.4 Limit Hahn ! Meixner. From the hyp ergeometric representation of the Hahn and Meixner p olynomials h ; n ( x; N ) = (  1) n ( N  1)!(  + n + 1) n !( N  n  1)!(  + 1) 3 F 2   x; +  + n +1 ;  n 1  N ; +1 ; 1  M  ; n ( x ) = (  ) n  n (   1) n 2 F 1   n;  x  ; 1  1   ; it is easy to check that the following limit relation holds lim N !1 h (1   )  N ;  1 n ( x; N ) = M  ; n ( x ) : (73) By using the well-known asymptotic formula for the  function (see for instance [1], Eq.(6.1.39) in page 257). ( aN + b )  p 2  e  aN ( aN ) aN + b  1 2 ; 17 and doing some straightforward, but tedious, calculation we obtain for the kernels K er H;; n  1 (0 ; 0) of the Hahn p olynomials the following expression in terms of the kernels of the Meixner ones lim N !1 K er H; (1   )  N ;  1 n  1 (0 ; 0) = K er M n  1 (0 ; 0) : From (54) we also notice that the constant  n;; A   n;; A; 0 of the representation formula (53) (here we are interested in the case when B = 0) is equal to lim N !1  n;; A = A  n (1   )  1 (  ) n n !(1 + AK er M n  1 (0 ; 0)) : From the last two expressions and taking into account Eqs. (73) and (42) we conclude that the following limit transition b etween Hahn and Meixner generalized p olynomials holds lim N !1 h (1   ) N  ;  1 ;A n ( x; N ) = M  ;;A n ( x ) : (74) 4.5 Limit Hahn ! Kravchuk. In a similar way, in this case we start from the classical relation lim t !1 h (1  p ) t;pt n ( x; N ) = K p n ( x; N  1) : (75) Notice that in this relation the Hahn p olynomials are dened for n < N , as well as the Kravchuk p olynomials are dened for n < N  1, i.e., the interval of orthogonality is reduced in one unit. Besides, for the kernels we have the expression lim t !1 K er H; (1  p ) t;pt n  1 (0 ; 0) = K er K n  1 (0 ; 0) ; and for the constant of the representation formula (53) lim t !1  n; (1  p ) t;pt A = A p n (1  p ) 1  n ( N  1)! n !( N  n  1)!(1 + AK er K n  1 (0 ; 0)) : Finally, using the last two expressions from (53) and (43) we obtain the limit relation lim t !1 h (1  p ) t;pt;A n ( x; N ) = K p;A n ( x; N  1) : (76) 4.6 Limit Hahn ! Jacobi. In this section we will analyze the limit relation involving Hahn and Jacobi p olynomials. As b efore we start from the classical relation lim N !1 2 n N n h ; n (( N  1) x; N ) = P ; n (2 x  1) : (77) In order to obtain the limit relation we will use the Eq. (49) for Hahn and Jacobi p olynomials. First of all, notice that lim N !1 K er H;; n  1 (0 ; 0) = K er J;; n  1 (  1 ;  1) ; lim N !1 K er H;; n  1 ( N  1 ; N  1) = K er J;; n  1 (1 ; 1) ; 18 h A;B ;; n ( x; N ) M  ;;A n K p;A n ( x ) C ;A n ( x ) P A;B ;; n ( x ) L ;A n ( x ) Hahn Meixner Jacobi Kravchuk Laguerre Charlier         + (78) ? (74) Q Q Q Q Q Q Q Q Qs (76) J J J J J J^ (80)       (68)       (70) @ @ @ @ @ @R (72) Figure 1: Limit relations involving the generalized p olynomials. and lim N !1 K er H;; n  1 (0 ; N  1) = K er J;; n  1 (  1 ; 1) : If we use now Eqs. (51), (52),(56) and (57), we conclude that lim N !1 2 n N n h A;B ;; n (0 ; N ) = P A;B ;; n (  1) ; and lim N !1 2 n N n h A;B ;; n ( N  1 ; N ) = P A;B ;; n ( 1) : The following limit relation b etween the norms of the Hahn ( d H n ) 2 and Jacobi ( d J n ) 2 p olynomials is also valid lim N !1 2 n N n ( d H n ) 2 = ( d J n ) 2 : Putting all these formulas in Eq. (49), taking the limit N ! 1 and using the classical relation (77) we nally obtain the limit relation b etween the generalized p olynomials, i.e., lim N !1 2 n N n h A;B ;; n (( N  1) x; N ) = P A;B ;; n (2 x  1) : (78) 19 4.7 Limit Jacobi ! Laguerre. Finally, we stablish the limit relation b etween Jacobi and Laguerre p olynomials. Like in the previous cases we start from the classical relation lim  !1 (  1) n  n 2 n P ; n  1  2 x   = L  n ( x ) : (79) From the last relation we notice that it is reasonable that the connection should b e b etween the Jacobi p olynomials with a mass p oint at x = 1 (i.e., A=0, B=A) and the generalized Laguerre p olynomials. In fact putting x = 0 we obtain lim  !1 (  1) n  n 2 n P ; n (1) = L  n (0) : Some straightforward calculations provide the relations lim  !1 K er J;; n  1 (1 ; 1) = K er L n  1 (0 ; 0) ; and for the norms of the Jacobi ( d J n ) 2 and Laguerre ( d L n ) 2 p olynomials lim  !1  2 n 2 2 n ( d J n ) 2 = ( d L n ) 2 : Then, from (39), (49) and (79) we obtain lim  !1 (  1) n  n 2 n P 0 ;A; n  1  2 x   = L ;A n ( x ) : (80) ACKNOWLEDGEMENTS The research of the authors was supp orted by Comision Interministerial de Ciencia y Tecnologa (CICYT) of Spain under grant PB 93-0228-C02-01 and by the INTAS. The authors thank to Professor Jessus Sanchez Dehesa (Universidad de Granada, Spain) for its useful suggestions and remarks and to the referees for they useful comments and remarks. References [1] M. Abramowitz and I.Stegun Editors: Handb o ok of Mathematical Functions. Dover Publications, Inc., New York, 1972. [2] R.  Alvarez-No darse and F. Marcellan: Dierence equation for modication of Meixner polynomials. J. Math. Anal. and Appl. 194 (3), (1995), 250-258. [3] R.  Alvarez-No darse, A.G. Garca and F. Marcellan: On the properties for modication of classical orthogonal polynomials of discrete variables. J. Comput. and Appl. Math. 65 , (1995) 3-18. [4] R.  Alvarez-No darse and F. Marcellan: The modication of classical Hahn polynomials of a discrete variable. Integral Transforms and Sp ecial Functions 4 (4), (1995), 243-262. [5] R.  Alvarez-No darse and F. Marcellan: A generalization of the classical Laguerre polynomials. Rend. Circ. Mat. Palermo. Serie I I, 44 , (1995), 315-329. 20 [6] R. Askey: Dierence equation for modications of Meixner polynomials. In Orthogonal Polynomials and their Applications. C. Brezinski et al. Editors. Annals on Computing and Applied Mathematics. J.C. Baltzer AG, Scientic Publishing Company, Basel. Volume 9, (1991), 418. [7] H. Bavinck and H. van Haeringen: Dierence equations for generalized Meixner polynomials. J. Math. Anal. and Appl. 184 (3), (1994), 453-463. [8] H. Bavinck and R. Ko eko ek: On a dierence equation for generalizations of Charlier polynomials. J. Approx. Th. 81 , (1995), 195-206. [9] T. S. Chihara: Orthogonal polynomials and measures with end point masses. Ro cky Mount. J. of Math., 15 (3),(1985),705-719. [10] T. S. Chihara: An intro duction to orthogonal p olynomials. Gordon and Breach. New York. 1978. [11] N. Dradi: Sur l'adjonction de deux masses de Dirac a une forme lineaire reguliere quelconque. Th  ese Doctorat de l'Universit  e Pierre et Marie Curie. Paris, 1990. [12] E. Go doy, F. Marcellan, L. Salto and A. Zarzo: Classical-type orthogonal polynomials: The discrete case. (manuscript submitted). [13] J. Ko eko ek and R.Ko eko ek : On a dierential equation for Koornwinder's generalized Laguerre polynomials. Pro c. Amer. Math. So c. 112 , (1991), 1045-1054. [14] R. Ko eko ek and H.G. Meijer: A generalization of Laguerre polynomials. SIAM J. Mat. Anal. 24 (3), (1993), 768-782. [15] R. Ko eko ek: Koornwinder's Laguerre Polynomials. Delft Progress Rep ort. 12 , (1988), 393-404. [16] R. Ko eko ek: Dierential equations for symmetric generalized ultraspherical polynomials. Transact. of the Amer. Math. So c. 345 , (1994), 47-72. [17] R. Ko eko ek: Generalizations of the classical Laguerre Polynomials and some qanalogues. Thesis, Delft University of Technology. 1990. [18] R. Ko eko ek and R.F. Swarttouw: The Askey-scheme of hypergeometric orthogonal polynomials and its q-analogue. Rep orts of the Faculty of Technical Mathematics and Informatics No. 94-05. Delft University of Technology, Delft, 1994. [19] T. H. Ko ornwinder: Orthogonal polynomials with weight function (1  x )  (1 + x )  + M  ( x + 1) + N  ( x  1) . Canad.Math.Bull, 27 (2), (1984), 205-214. [20] H. L. Krall: On Orthogonal Polynomials satisfying a certain fourth order dierential equation. The Pennsylvania State College Bulletin, 6 , (1940), 1-24. [21] H.K. Kwon and S.B. Park: Two points masses perturbation of regular moment functionals. Indag. Math. (In press) [22] F. Marcellan and P. Maroni: Sur l'adjonction d'une masse de Dirac a une forme reguliere et semi-classique. Ann.Mat.Pura ed Appl., IV , CLXI I, (1992), 1-22. [23] A.F.Nikiforov, S.K.Suslov and V.B.Uvarov: Orthogonal Polynomials in Discrete Variables. Springer Series in Computational Physics. Springer-Verlag,Berlin, 1991. 21 [24] E.D. Rainville: Sp ecial Functions. Chelsea Publishing Company, New York, 1971. [25] G. Szego: Orthogonal Polynomials. Amer. Math. So c. Collo q. Publ., 23 , Amer. Math. So c., Providence, RI, 1975 (4th edition). [26] S. Wolfram: MATHEMATICA . A system for doing Mathematics by Computer . Addison-Wesley Publishing Co., New York, 1991. [27] Wolfram Research Inc. Guide to Standard Mathematica Packages , version 2.2, Wolfram Research, Inc., Champaign, Illinois, (1993). 22