Selçuk J. Appl. Math. Selçuk Journal of Vol. 8. No. 1. pp. 9-14 , 2007 Applied Mathematics
On the norms of Hankel matrices with the Motzkin numbers Aynur Yalçıner and Necati Ta¸skara
Department of Mathematics, Selcuk University, 42031 (Campus) Konya, Turkey e-mail:ayalciner@ selcuk.edu.tr, ntaskara@ selcuk.edu.tr
Received: November 11, 2006
Summary:We give bounds for the norm of the Hadamard inverse of Hankel
matrices with Motzkin numbers.Furthermore, we determine an upper bound for the Hadamard product of these matrices.
Key words: Motzkin number; Hadamard inverse; Hankel matrix; norm;
Euclidean norm. 1. Introduction
The Motzkin numbers are defined by the recursion
= −1+ X−2 =0
−2− 0= 1
with generating functions
() =³1 − −p1 − 2 − 32´22
The first few numbers are thus
0 1 2 3 4 5 6 7
1 1 2 4 9 21 51 127
It was shown in [1] that, for the Motzkin numbers, (i) −1 ≤ +1 ( ≥ 1) (ii) −1 3 ( ≥ 1) (iii) lim →∞ −1 = 3.
(1.1) = ⎡ ⎢ ⎢ ⎢ ⎣ 0 1 · · · 1 2 · · · +1 .. . +1 · · · 2 ⎤ ⎥ ⎥ ⎥ ⎦ and (1.2) = ⎡ ⎢ ⎢ ⎢ ⎣ 1 2 · · · 2 3 · · · +1 .. . +1 · · · 2−1 ⎤ ⎥ ⎥ ⎥ ⎦
and obtained following proposition: Proposition (i)det = 1 for all ,
(ii) det = 1, 0, −1 for ≡ 0, 1 (mod 6), ≡ 2, 5 (mod 6) and ≡ 3, 4
(mod 6) respectively [1].
In this study, we have obtained an upper bound for the norm of the Hadamard
inverse of Hankel matrix in the following form:
(1.3) = ⎡ ⎢ ⎢ ⎢ ⎣ 2 3 · · · +1 3 4 · · · +2 .. . +1 +2 · · · 2 ⎤ ⎥ ⎥ ⎥ ⎦
and we have found an upper bound for the Euclidean norm of Hadamard product of these Hankel matrices.
Now, we start with some preliminaries. Let = () and = () be ×
matrices. The Hadamard product of and is defined by ◦ = ()
The Hadamard inverse of (with 0, 1 ≤ ≤ , 1 ≤ ≤ ) is defined by
◦(−1)=³ 1
´ [6].
Let be any × matrix. The norm of the matrix is
kk= ⎛ ⎝ X =1 X =1 || ⎞ ⎠ 1 (1 ≤ ∞) If = ∞ then kk∞= lim →∞kk= max ||
kk = ⎛ ⎝ X =1 X =1 ||2 ⎞ ⎠ 12
and also the spectral norm is kk2=
r max
1≤≤
()
where is × and is the conjugate transpose of the matrix . A function is called a Riemann-zeta function if
() = P∞
=1
1
( 1)
2. Main Results
Theorem 1Let the matrix be as in (1.3). Then
° ° °◦(−1) ° ° ° ≤ s () + 1 4 where 1 ∞.
Proof.From the definition of norm, we have
° ° °◦(−1) ° ° ° = X =1 +1 + X−1 =1 − ++1 For the first term, we have
(2.1) X =1 +1 X =1 1
Let us prove (2.1) by induction. For = 1 inequality (2.1) is true. We shall assume the result is true for − 1 ∈ N. Since ∈ N and +1≥ 2 we have
+1
1
Hence for ∈ N inequality (2.1) is true. We consider the limit case as → ∞ in (2.1) P =1 +1 () On the other hand, for ≥ 1, we write
(2.2) +1 ≥ 2
We can prove (2.2) by induction. For = 1 inequality (2.2) is true. We shall assume that for − 1 ∈ N,
−1 ≥ 2 is true. Since −1 ≤ +1 we have +1 ≥ 2
Consequently for ∈ N inequality (2.2) is true. It is easy to see that for ≥ 4
X−1 =1 − ++1 ≤ 1 4
Using inequality (2.2), we have ++1 4+−3 and therefore, we write
−1 X =1 − ++1 1 4 X−1 =1 − + −3 = 1 4 µ − 1 −2 + − 2 −1 + + 1 2 −4 ¶ Since −1 −2 −2 −1 1 2−4 we obtain −1 X =1 − ++1 1 4 ( − 1)2 −2 For ≥ 5, since −2 ≥ ( − 1)2and 0
(−1)2 −2 1 we have (2.3) X−1 =1 − ++1 ≤ 1 4 Thus the proof is completed.
Corollary 1Let the matrix be as (1.1). Then
° ° °◦(−1) ° ° °≤ s 3 + () + 1 2 where 1 ∞.
Corollary 2Let the matrix be as (1.2). Then
° ° °◦(−1) ° ° °≤ s 1 + () + 1 3
where 1 ∞ .
Theorem 2Let and matrices as in (1.3) and (1.2), respectively.Then
° ° °◦ ◦(−1) ° ° ° ≤ 3
where k·k is the Euclidean norm.
Proof.From the definition of Euclidean norm, we have ° ° °◦ ◦(−1) ° ° °2 = X =1 µ +1 ¶2 + −1 X =1 ( − ) µ ++1 + ¶2 since +1 3 ( ≥ 1), we obtain ° ° °◦ ◦(−1) ° ° °2 = 9 X =1 + 9 X−1 =1 ( − ) = 92 Hence, ° ° °◦ ◦(−1) ° ° ° ≤ 3
Example 1Let = 2. For inequality (2.1) , we have the following values: n P =1 +1 P =1 1 2 3 4 5 10 20 30 40 50 100 0.25000000 0.37500000 0.41203704 0.42110735 0.42348212 0.42348248 0.42348248 0.42348248 0.42348248 0.42348248 1.00000000 1.25000000 1.36111116 1.42361116 1.53976774 1.59366302 1.61103916 1.61961925 1.62473297 1.63488400
n P− =1 − ++1 1 4 2 3 4 5 6 7 8 9 10 11 12 0.01234568 0.00491962 0.00128699 0.00027983 0.00005435 0.00000978 0.00000166 0.00000027 0.00000004 0.00000001 0.00000000 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 0.01234568 References
1.M. Aigner, Motzkin numbers, Europ. J. Combinatorics 19(1998), 663-675. 2.A. Yalçıner, Hankel ve Circulant Matrislerin Terslerinin Normları Üzerine, (2006), Doktora Tezi,Fen Bilimleri Enstitüsü, Konya.
3. R. Donaghey, L.W. Shapiro, Motzkin numbers, Journal of Combinatorial Theory, Series A 23 (1977), 291-301.
4.R.Stanley, Enumerative Combinatorics, Vol II, Cambridge Univ. Press, 1999. 5.R.A.Horn, C.R. Johnson, Topics in Matrix Analysis, Cambridge University Press, 1991.
6. R.A.Horn, The Hadamard product, in: C.R. Johnson (Ed.), Proc. Appl. Math., vol. 40. American Mathematical Society, Providence, Rhode Island. (1990) 87-169.