calculations on some sequence spaces - European Mathematical ...
calculations on some sequence spaces - European Mathematical ...
calculations on some sequence spaces - European Mathematical ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
IJMMS 2004:31, 1653–1670<br />
PII. S0161171204209401<br />
http://ijmms.hindawi.com<br />
© Hindawi Publishing Corp.<br />
CALCULATIONS ON SOME SEQUENCE SPACES<br />
BRUNO DE MALAFOSSE<br />
Received 16 September 2002<br />
We deal with space of <strong>sequence</strong>s generalizing the well-known <strong>spaces</strong> w p ∞(λ), c ∞ (λ,µ), replacing<br />
the operators C(λ) and ∆(µ) by their transposes. We get generalizati<strong>on</strong>s of results<br />
c<strong>on</strong>cerning the str<strong>on</strong>g matrix domain of an infinite matrix A.<br />
2000 Mathematics Subject Classificati<strong>on</strong>: 46A45, 40C05.<br />
1. Notati<strong>on</strong>s and preliminary results. For a given infinite matrix A = (a nm ) n,m≥1 ,<br />
the operators A n are defined, for any integer n ≥ 1, by<br />
A n (X) =<br />
∞∑<br />
a nm x m , (1.1)<br />
m=1<br />
where X = (x n ) n≥1 , the series intervening in the sec<strong>on</strong>d member being c<strong>on</strong>vergent. So<br />
we are led to the study of the infinite linear system<br />
A n (X) = b n , n= 1,2,..., (1.2)<br />
where B = (b n ) n≥1 is a <strong>on</strong>e-column matrix and X the unknown, see [1, 2, 3, 4, 5, 6, 7,<br />
8, 10]. Equati<strong>on</strong> (1.2) can be written in the form AX = B, where AX = (A n (X)) n≥1 .In<br />
this paper, we will also c<strong>on</strong>sider A an operator from a <strong>sequence</strong> space into another<br />
<strong>sequence</strong> space.<br />
A Banach space E of complex <strong>sequence</strong>s with the norm ‖‖ E is a BK space if each<br />
projecti<strong>on</strong> P n X = x n is c<strong>on</strong>tinuous for all X ∈ E. A BK space E is said to have AK, (see<br />
[12, 13]), if B = ∑ ∞<br />
m=1 b m e m , for every B = (b n ) n≥1 ∈ E, (withe n = (0,...,1,...),1being<br />
in the nth positi<strong>on</strong>), that is,<br />
∥<br />
∞∑<br />
m=N+1<br />
b m e m<br />
∥ ∥∥∥∥E<br />
→ 0 (n →∞). (1.3)<br />
We will write s for the set of all complex <strong>sequence</strong>s, l ∞ , c, c 0 for the sets of bounded,<br />
c<strong>on</strong>vergent, and null <strong>sequence</strong>s, respectively. We will denote by cs and l 1 the sets of<br />
c<strong>on</strong>vergent and absolutely c<strong>on</strong>vergent series, respectively.<br />
In all that follows we will use the set<br />
U +∗ = {( u n<br />
)n≥1 ∈ s | u n > 0 ∀n } . (1.4)
1654 BRUNO DE MALAFOSSE<br />
From Wilansky’s notati<strong>on</strong>s [15], we define for any <strong>sequence</strong><br />
α = ( α n<br />
)n≥1 ∈ U +∗ , (1.5)<br />
and for any set of <strong>sequence</strong>s E, the set<br />
( 1<br />
α<br />
) −1 { ∣ ( }<br />
(xn ) ∣∣∣<br />
∗E =<br />
n≥1 ∈ s xn<br />
∈ E . (1.6)<br />
α n<br />
)n<br />
We will write α∗E instead of (1/α) −1 ∗E forshort.Soweget<br />
⎧<br />
s ⎪⎨<br />
α ◦ if E = c 0 ,<br />
α∗E = s α (c) if E = c,<br />
⎪⎩<br />
s α if E = l ∞ .<br />
(1.7)<br />
We have for instance<br />
α∗c 0 = s ◦ α = {( x n<br />
)n≥1 ∈ s | x n = o ( α n<br />
)<br />
n →∞<br />
}<br />
. (1.8)<br />
Each of the <strong>spaces</strong> α∗E, where E ∈{c 0 ,c,l ∞ }, is a BK space normed by<br />
( ∣ ∣ ) ∣xn ‖X‖ sα = sup , (1.9)<br />
n≥1 α n<br />
and s ◦ α has AK.<br />
Now let α = (α n ) n≥1 and β = (β n ) n≥1 ∈ U +∗ . S α,β is the set of infinite matrices<br />
A = (a nm ) n,m≥1 such that<br />
(<br />
anm α m<br />
)<br />
m≥1 ∈ l1 ∀n ≥ 1,<br />
S α,β is a Banach space with the norm<br />
∞∑ (∣ ∣ ) ( )<br />
∣anm∣αm = O βn<br />
m=1<br />
(n →∞). (1.10)<br />
⎛<br />
‖A‖ Sα,β = sup⎝<br />
n≥1<br />
⎞<br />
∞∑<br />
∣<br />
∣ anm α m ⎠ . (1.11)<br />
β n<br />
m=1<br />
Let E and F be any subsets of s.WhenA maps E into F,wewillwriteA ∈ (E,F),see[11].<br />
So for every X ∈ E, AX ∈ F, (AX ∈ F will mean that for each n ≥ 1 the series defined<br />
by y n = ∑ ∞<br />
m=1 a nm x m is c<strong>on</strong>vergent and (y n ) n≥1 ∈ F). It has been proved in [9] that<br />
A ∈ (s α ,s β ) if and <strong>on</strong>ly if A ∈ S α,β . So we can write that (s α ,s β ) = S α,β .<br />
When s α = s β , we obtain the unital Banach algebra S α,β = S α ,(see[1, 2, 3, 5, 6, 10])<br />
normed by ‖A‖ Sα =‖A‖ Sα,α .<br />
We also have A ∈ (s α ,s α ) if and <strong>on</strong>ly if A ∈ S α .If‖I −A‖ Sα < 1, we will say that A ∈ Γ α .<br />
Since the set S α is a unital algebra, we have the useful result that if A ∈ Γ α , A is bijective<br />
from s α into itself.
CALCULATIONS ON SOME SEQUENCE SPACES 1655<br />
If α = (r n ) n≥1 , Γ α , S α , s α , sα ◦ ,ands(c) α are replaced by Γ r , S r , s r , sr ◦ , and s(c) r , respectively,<br />
(see [1, 2, 3, 5, 6, 10]). When r = 1, we obtain s 1 = l ∞ , s1 ◦ = c 0, and s (c)<br />
1 = c, and<br />
putting e = (1,1,...), we have S 1 = S e . It is well known, see [11], that<br />
(<br />
s1 ,s 1<br />
)<br />
=<br />
(<br />
c0 ,s 1<br />
)<br />
=<br />
(<br />
c,s1<br />
)<br />
= S1 . (1.12)<br />
For any subset E of s, weput<br />
AE ={Y ∈ s |∃X ∈ E, Y = AX}. (1.13)<br />
If F is a subset of s, wewilldenote<br />
F(A)= F A ={X ∈ s | Y = AX ∈ F}. (1.14)<br />
We can see that F(A)= A −1 F.<br />
2. Some properties of the operators ∆ + and Σ + . Here we will deal with the operators<br />
represented by C + (λ) and ∆ + (λ).<br />
Let<br />
U = {( u n<br />
)n≥1 ∈ s | u n ≠ 0 ∀n } . (2.1)<br />
We define C(λ) = (c nm ) n,m≥1 ,forλ = (λ n ) n≥1 ∈ U, by<br />
⎧<br />
⎪⎨<br />
1<br />
if m ≤ n,<br />
c nm = λ n<br />
⎪⎩<br />
0 otherwise.<br />
(2.2)<br />
So, we put C + (λ) = C(λ) t . It can be proved that the matrix ∆(λ) = (c ′ nm) n,m≥1 with<br />
⎧<br />
λ ⎪⎨ n if m = n,<br />
c nm ′ = −λ n−1 if m = n−1, n≥ 2,<br />
⎪⎩<br />
0 otherwise,<br />
(2.3)<br />
is the inverse of C(λ),see[12, 14]. Similarly, we put ∆ + (λ) = ∆(λ) t .Ifλ = e, weget<br />
the well-known operator of first difference represented by ∆(e) = ∆ and it is usually<br />
written as Σ = C(e). Note that ∆ = Σ −1 and Σ bel<strong>on</strong>g to any given space S R with R>1.<br />
Writing D λ = (λ n δ nm ) n,m≥1 , (where δ nm = 0forn ≠ m and δ nn = 1 otherwise), we have<br />
∆ + (λ) = D λ ∆ + .Soforanygivenα ∈ U +∗ , we see that if (α n−1 /α n )|λ n /λ n−1 |=O(1),<br />
then ∆ + (λ) ∈ (s (α/|λ|) ,s α ).SinceKer∆ + (λ) ≠ 0, we are lead to define the set<br />
sα<br />
∗ (<br />
∆ + (λ) ) (<br />
= s α ∆ + (λ) ) ⋂<br />
s(α/|λ|) = { X = ( )<br />
x n n≥1 ∈ s (α/|λ|) | ∆ + }<br />
(λ)X ∈ s α . (2.4)
1656 BRUNO DE MALAFOSSE<br />
It can easily be seen that<br />
s ∗ (α/|λ|)(<br />
∆ + (e) ) = s ∗ (α/|λ|)(<br />
∆<br />
+ ) = s ∗ α<br />
(<br />
∆ + (λ) ) . (2.5)<br />
2.1. Properties of the <strong>sequence</strong> C(α)α. We will use the following sets:<br />
⎧<br />
⎛<br />
⎨ ∣ ∣∣∣<br />
Ĉ 1 =<br />
⎩ α ∈ U +∗ 1<br />
n∑<br />
⎝<br />
α n<br />
k=1<br />
⎧<br />
⎛<br />
⎨ ∣ ∣∣∣<br />
Ĉ =<br />
⎩ α ∈ U +∗ 1<br />
n∑<br />
⎝<br />
α n<br />
⎧<br />
⎨<br />
Ĉ 1 + = ⎩ α ∈ U ⋂ +∗ cs<br />
∣<br />
k=1<br />
1<br />
α n<br />
⎛<br />
⎝<br />
α k<br />
⎞<br />
⎠ = O(1) (n →∞)<br />
⎫<br />
⎬<br />
⎭ ,<br />
α k<br />
⎞<br />
⎠ ∈ c<br />
⎫<br />
⎬<br />
⎭ ,<br />
∞∑<br />
k=n<br />
{ ∣ ( ) }<br />
∣∣∣<br />
Γ = α ∈ U +∗ αn−1<br />
lim < 1 ,<br />
n→∞ α n<br />
{ ∣ ( ) }<br />
∣∣∣<br />
Γ + = α ∈ U +∗ αn+1<br />
lim < 1 .<br />
n→∞ α n<br />
α k<br />
⎞<br />
⎠ = O(1) (n →∞)<br />
⎫<br />
⎬<br />
⎭ ,<br />
(2.6)<br />
Note that α ∈ Γ + if and <strong>on</strong>ly if 1/α ∈ Γ . We will see in Propositi<strong>on</strong> 2.1 that if α ∈ Ĉ 1 , α<br />
tends to infinity. On the other hand, we see that ∆ ∈ Γ α implies α ∈ Γ and α ∈ Γ if and<br />
<strong>on</strong>ly if there is an integer q ≥ 1 such that<br />
( )<br />
αn−1<br />
γ q (α) = sup < 1. (2.7)<br />
n≥q+1 α n<br />
We obtain the following results in which we put [C(α)α] n = ( ∑ n<br />
k=1 α k)/α n .<br />
Propositi<strong>on</strong> 2.1. Let α ∈ U +∗ .Then<br />
(i) α n−1 /α n → 0 if and <strong>on</strong>ly if [C(α)α] n → 1,<br />
(ii) (a) α ∈ Ĉ implies that (α n−1 /α n ) n≥1 ∈ c,<br />
(b) [C(α)α] n → l implies that α n−1 /α n → 1−1/l,<br />
(iii) if α ∈ Ĉ 1 , there are K>0 and γ>1 such that<br />
α n ≥ Kγ n ∀n, (2.8)<br />
(iv) the c<strong>on</strong>diti<strong>on</strong> α ∈ Γ implies that α ∈ Ĉ 1 and there exists a real b>0 such that<br />
[<br />
C(α)α<br />
]n ≤ 1<br />
1−χ +bχn for n ≥ q +1, χ= γ q (α) ∈ ]0,1[, (2.9)<br />
(v) the c<strong>on</strong>diti<strong>on</strong> α ∈ Γ + implies that α ∈ Ĉ+ 1 .
CALCULATIONS ON SOME SEQUENCE SPACES 1657<br />
Proof.<br />
Assume that α n−1 /α n → 0. Then there is an integer N such that<br />
n ≥ N +1 ⇒ α n−1<br />
α n<br />
≤ 1 2 . (2.10)<br />
So there exists a real K>0 such that α n ≥ K2 n for all n and<br />
Then<br />
⎛<br />
n−1<br />
1 ∑<br />
⎝<br />
α n<br />
k=1<br />
α k<br />
=<br />
α ( )<br />
k<br />
··· αn−1 1 n−k<br />
≤ for N ≤ k ≤ n−1. (2.11)<br />
α n α k+1 α n 2<br />
⎞ ⎛<br />
N−1<br />
α k<br />
⎠<br />
1 ∑<br />
= ⎝<br />
α n<br />
k=1<br />
⎞<br />
n−1<br />
∑<br />
α k<br />
⎠ +<br />
k=N<br />
⎛<br />
α k<br />
≤ 1<br />
N−1<br />
∑<br />
⎝<br />
α n K2 n<br />
k=1<br />
⎞<br />
n−1<br />
∑ ( )<br />
α k<br />
⎠ 1 n−k<br />
+ , (2.12)<br />
2<br />
k=N<br />
and since ∑ n−1<br />
k=N(1/2) n−k = 1−(1/2) n−N → 1, (n →∞), wededucethat<br />
⎛ ⎞<br />
n−1<br />
1 ∑<br />
⎝ α k<br />
⎠ = O(1) (2.13)<br />
α n<br />
k=1<br />
and ([C(α)α] n ) ∈ l ∞ . Using the identity<br />
[<br />
C(α)α<br />
]n = α 1 +···+α n−1<br />
α n−1<br />
α n−1<br />
α n<br />
we get [C(α)α] n → 1. This proves the necessity.<br />
C<strong>on</strong>versely, if [C(α)α] n → 1, then<br />
[ ]<br />
α n−1 C(α)α<br />
n<br />
= [ ]<br />
−1<br />
α n C(α)α n−1<br />
+1 = [ C(α)α ] n−1<br />
(<br />
αn−1<br />
)<br />
+1, (2.14)<br />
α n<br />
→ 0. (2.15)<br />
(ii) is a direct c<strong>on</strong><strong>sequence</strong> of the identity (2.14).<br />
(iii) We put Σ n = ∑ n<br />
k=1 α k .ThenforarealM>1,<br />
[<br />
C(α)α<br />
]n = Σ n<br />
Σ n −Σ n−1<br />
≤ M ∀n. (2.16)<br />
So Σ n ≥ (M/(M −1))Σ n−1 and Σ n ≥ ( M/(M−1) ) n−1<br />
α1 ∀n. Therefore, from<br />
( )<br />
α 1 M n−1<br />
≤ [ C(α)α ] n<br />
α n M −1<br />
= Σ n<br />
≤ M, (2.17)<br />
α n<br />
we c<strong>on</strong>clude that α n ≥ Kγ n for all n, withK = (M −1)α 1 /M 2 and γ = M/(M−1)>1.
1658 BRUNO DE MALAFOSSE<br />
(iv) If α ∈ Γ , there is an integer q ≥ 1 for which<br />
So there is a real M ′ > 0 for which<br />
k ≥ q +1 implies α k−1<br />
α k<br />
≤ χ
CALCULATIONS ON SOME SEQUENCE SPACES 1659<br />
X = (x n,0 ) n≥1 in s α . Then there is a unique scalar x 1 such that<br />
n−1<br />
∑<br />
x n,0 = x 1 − α k . (2.25)<br />
k=1<br />
So<br />
∣ ⎛<br />
⎞<br />
∣ xni , 0 n i −1<br />
=<br />
1 ∑<br />
⎝<br />
α x1 − α k<br />
⎠<br />
ni ∣ α →∞ as i →∞, (2.26)<br />
ni ∣<br />
k=1<br />
and X ∉ s α , which is c<strong>on</strong>tradictory.<br />
We c<strong>on</strong>clude that each of the properties e ∈ Ker∆ + ⋂ s α and e ∉ Ker∆ + ⋂ s α is impossible<br />
and ∆ + is not bijective from s α into itself. This proves the lemma.<br />
Lemma 2.6. For every X ∈ c 0 , Σ + (∆ + X) = X and for every X ∈ cs, ∆ + (Σ + X) = X.<br />
Proof.<br />
It can easily be seen that<br />
[<br />
Σ<br />
+ ( ∆ + X )] n = ∞ ∑<br />
(<br />
xm −x m+1<br />
)<br />
= xn ∀X ∈ c 0 ,<br />
m=n<br />
[<br />
∆<br />
+ ( Σ + X )] ∞ n = ∑<br />
∞∑<br />
x m − x m = x n<br />
m=n m=n+1<br />
∀X ∈ cs.<br />
(2.27)<br />
We can assert the following result, in which we put α + = (α n+1 ) n≥1 and s ◦∗<br />
α (∆ + ) =<br />
s ◦ α(∆ + ) ⋂ s ◦ α. Note that from (2.5) we have<br />
s ∗ α<br />
(<br />
∆ + (e) ) = s ∗ α<br />
(<br />
∆<br />
+ ) = s α<br />
(<br />
∆<br />
+ ) ⋂<br />
sα . (2.28)<br />
Theorem 2.7. (i) (a) s α (∆) = s α if and <strong>on</strong>ly if α ∈ Ĉ 1 ,<br />
(b) sα(∆) ◦ = sα ◦ if and <strong>on</strong>ly if α ∈ Ĉ 1 ,<br />
(c) s α<br />
(c) (∆) = s α (c) if and <strong>on</strong>ly if α ∈ Ĉ.<br />
(ii) (a) α ∈ Ĉ 1 if and <strong>on</strong>ly if s α +(∆ + ) = s α and ∆ + is surjective from s α into s α +,<br />
(b) α ∈ Ĉ+ 1 if and <strong>on</strong>ly if s∗ α (∆ + ) = s α and ∆ + is bijective from s α into s α ,<br />
(c) α ∈ Ĉ+ 1 implies that s◦∗ α (∆ + ) = sα ◦ and ∆ + is bijective from sα ◦ into sα.<br />
◦<br />
(iii) α ∈ Ĉ+ 1 if and <strong>on</strong>ly if s α(Σ + ) = s α and s α (Σ + ) = s α implies sα ◦ (Σ+ ) = sα ◦ .<br />
Proof. (i) has been proved in [8].<br />
(ii)(a) Sufficiency. If ∆ + is surjective from s α into s α +, then for every B ∈ s α +<br />
soluti<strong>on</strong>s of ∆ + X = B in s α are given by<br />
the<br />
n∑<br />
x n+1 = x 1 − b k n = 1,2,..., (2.29)<br />
k=1
1660 BRUNO DE MALAFOSSE<br />
where x 1 is arbitrary. If we take B = α + ,wegetx n = x 1 − ∑ n<br />
k=2 α k.So<br />
⎛<br />
x n<br />
= x 1<br />
− 1 ⎝<br />
α n α n α n<br />
n∑<br />
k=2<br />
α k<br />
⎞<br />
⎠ = O(1). (2.30)<br />
Taking x 1 =−α 1 , we c<strong>on</strong>clude that ( ∑ n−1<br />
k=1 α k )/α n = O(1) and α ∈ Ĉ 1 .<br />
C<strong>on</strong>versely, assume that α ∈ Ĉ 1 . From the inequality<br />
α n−1<br />
α n<br />
≤ 1 α n<br />
⎛<br />
⎝<br />
n∑<br />
k=1<br />
α k<br />
⎞<br />
⎠ = O(1), (2.31)<br />
we deduce that α n−1 /α n = O(1) and ∆ + ∈ (s α ,s α +). ThenforanygivenB ∈ s α +,the<br />
soluti<strong>on</strong>s of the equati<strong>on</strong> ∆ + X = B are given by x 1 =−u and<br />
n−1<br />
∑<br />
−x n = u+ b k for n ≥ 2, (2.32)<br />
where u is an arbitrary scalar. So there exists a real K>0 such that<br />
∣ ∣ xn ∣<br />
∑ n−1 u+<br />
=<br />
α n<br />
k=1 b k<br />
α n<br />
∣<br />
k=1<br />
≤ |u|+K(∑ n<br />
k=2 α )<br />
k<br />
= O(1) (2.33)<br />
α n<br />
and X ∈ s α . We c<strong>on</strong>clude that ∆ + is surjective from s α into s α +.<br />
(ii)(b) Necessity. Assume that α ∈ Ĉ+ 1 .Then∆+ ∈ (s α ,s α ),since<br />
α n+1<br />
α n<br />
≤ 1 α n<br />
⎛<br />
⎝<br />
∞∑<br />
k=n<br />
α k<br />
⎞<br />
⎠ = O(1) (n →∞). (2.34)<br />
Further, from s α ⊂ cs, we deduce, using Lemma 2.4, that for any given B ∈ s α ,<br />
∆ +( Σ + B ) = B. (2.35)<br />
On the other hand, Σ + B = ( ∑ ∞<br />
k=n b k ) n≥1 ∈ s α ,sinceα ∈ Ĉ+ 1 .So∆+ is surjective from s α<br />
into s α . Finally, ∆ + is injective because the equati<strong>on</strong><br />
∆ + X = O (2.36)<br />
admits the unique soluti<strong>on</strong> X = O in s α ,since<br />
Ker∆ + = { ue t | u ∈ C } (2.37)<br />
and e t ∉ s α .
CALCULATIONS ON SOME SEQUENCE SPACES 1661<br />
Sufficiency. For every B ∈ s α , the equati<strong>on</strong> ∆ + X = B admits a unique soluti<strong>on</strong> in<br />
s α .ThenfromLemma 2.5, α ∈ cs and since s α ⊂ cs, we deduce from Lemma 2.6 that<br />
X = Σ + B ∈ s α is the unique soluti<strong>on</strong> of ∆ + X = B. TakingB = α, wegetΣ + α ∈ s α ,thatis,<br />
α ∈ Ĉ+ 1 .<br />
(ii)(c) If α ∈ Ĉ+ 1 , ∆+ is bijective from sα ◦ into itself. Indeed, we have D 1/α ∆ + D α ∈ (c 0 ,c 0 )<br />
from (2.34) andLemma 2.4. Furthermore, since α ∈ Ĉ+ 1 we have s◦ α ⊂ cs and for every<br />
B ∈ sα,<br />
◦<br />
∆ +( Σ + B ) = B. (2.38)<br />
From Lemma 2.4, we have Σ + ∈ (sα,s ◦ α), ◦ so the equati<strong>on</strong> ∆ + X = B admits the soluti<strong>on</strong><br />
X 0 = Σ + B in sα ◦ and we have proved that ∆ + is surjective from sα ◦ into itself. Finally,<br />
α ∈ Ĉ+ 1 implies that et ∉ sα,soKer∆ ⋂ ◦ + sα ◦ ={0} and we c<strong>on</strong>clude that ∆ + is bijective<br />
from sα ◦ into itself.<br />
(iii) comes from (ii), since α ∈ Ĉ+ 1 if and <strong>on</strong>ly if ∆+ is bijective from s α into itself and<br />
Σ +( ∆ + X ) = ∆ +( Σ + X ) = X ∀X ∈ s α . (2.39)<br />
As a direct c<strong>on</strong><strong>sequence</strong> of Theorem 2.7 we obtain the following results.<br />
Corollary 2.8.<br />
Let R be any real > 0. Then<br />
R>1 ⇐⇒ s R (∆) = s R ⇐⇒ s ◦ R (∆) = s◦ R ⇐⇒ s R(<br />
∆<br />
+ ) = s R . (2.40)<br />
Proof. From (i) and (ii) in Theorem 2.7, we see that it is enough to prove that α =<br />
(R n ) n≥1 ∈ Ĉ 1 if and <strong>on</strong>ly if R>1. We have (R n ) n≥1 ∈ Ĉ 1 if and <strong>on</strong>ly if R ≠ 1 and<br />
R −n ⎛<br />
⎝<br />
n∑<br />
k=1<br />
⎞<br />
R k ⎠<br />
1 =<br />
1−R R−n+1 −<br />
This means that R>1 and the corollary is proved.<br />
R = O(1)<br />
1−R<br />
as n →∞. (2.41)<br />
Using the notati<strong>on</strong> α − = (1,α 1 ,α 2 ,...,α n−1 ,...) we get the next result.<br />
Corollary 2.9. Let α ∈ U +∗ and µ ∈ U. Then<br />
(i) α/|µ|∈Ĉ 1 if and <strong>on</strong>ly if<br />
s α<br />
(<br />
∆ + (µ) ) = s (α/|µ|) −, (2.42)<br />
(ii) α/|µ|∈Ĉ+ 1<br />
if and <strong>on</strong>ly if sα<br />
∗ (<br />
∆ + (µ) ) = s (α/|µ|) . (2.43)<br />
Proof.<br />
First we have<br />
s α<br />
(<br />
∆ + (µ) ) = s (α/|µ|)<br />
(<br />
∆<br />
+ ) . (2.44)
1662 BRUNO DE MALAFOSSE<br />
Indeed,<br />
X ∈ s α<br />
(<br />
∆ + (µ) ) ⇐⇒ D µ ∆ + X ∈ s α ⇐⇒ ∆ + X ∈ s (α/|µ|) ⇐⇒ X ∈ s (α/|µ|)<br />
(<br />
∆<br />
+ ) . (2.45)<br />
Now, if α/|µ| ∈ Ĉ 1 , from (i) in Theorem 2.7, we have s (α/|µ|) (∆ + ) = s (α/|µ|) − and<br />
s α (∆ + (µ)) = s (α/|µ|) −. C<strong>on</strong>versely, assume s α (∆ + (µ)) = s (α/|µ|) −. Reas<strong>on</strong>ing as above,<br />
we get s (α/|µ|) (∆ + ) = s (α/|µ|) −, and using (i) in Theorem 2.7 we c<strong>on</strong>clude that α/|µ|∈Ĉ 1<br />
and (i) holds.<br />
(ii) α/|µ|∈Ĉ+ 1 implies that ∆+ is bijective from s (α/|µ|) into itself. Thus<br />
s ∗ α<br />
(<br />
∆ + (µ) ) = s ∗ (α/|µ|)(<br />
∆<br />
+ ) = s (α/|µ|) . (2.46)<br />
This proves the necessity. C<strong>on</strong>versely, assume that sα ∗ (∆ + (µ))=s (α/|µ|) .Thens ∗ (α/|µ|) (∆+ )<br />
= s (α/|µ|) and from Theorem 2.7(ii)(b), α/|µ|∈Ĉ+ 1 and (ii) holds.<br />
2.3. Spaces w p α(λ) and w +p<br />
α (λ) for given real p>0. Here we will define sets generalizing<br />
the well-known sets<br />
w p ∞(λ) = { X ∈ s | C(λ) ( |X| p) ∈ l ∞<br />
}<br />
,<br />
w p 0 (λ) = { X ∈ s | C(λ) ( |X| p) ∈ c 0<br />
}<br />
,<br />
(2.47)<br />
see [9, 12, 13, 14, 15]. It is proved that each of the sets w p 0 = wp 0 ((n) n) and w∞ p =<br />
w∞((n) p n ) is a p-normed FK space for 0
CALCULATIONS ON SOME SEQUENCE SPACES 1663<br />
Theorem 2.10. (i) (a) The c<strong>on</strong>diti<strong>on</strong> α ∈ Ĉ+ 1<br />
is equivalent to<br />
w +p<br />
α (λ) = s (αλ) 1/p . (2.51)<br />
(b) If α ∈ Ĉ+ 1 , then<br />
w ◦p<br />
α (λ) = s ◦ (αλ) 1/p . (2.52)<br />
(ii) (a) The c<strong>on</strong>diti<strong>on</strong> αλ ∈ Ĉ 1 is equivalent to<br />
w p α(λ) = s (αλ) 1/p . (2.53)<br />
(b) If αλ ∈ Ĉ 1 , then<br />
w ◦+p<br />
α (λ) = s ◦ (αλ) 1/p . (2.54)<br />
Proof.<br />
Assume that α ∈ Ĉ+ 1 .SinceC+ (λ) = Σ + D 1/λ , we have<br />
w +p<br />
α (λ) = { X | ( Σ + D 1/λ<br />
)(<br />
|X|<br />
p ) ∈ s α<br />
}<br />
=<br />
{<br />
X | D1/λ<br />
(<br />
|X|<br />
p ) ∈ s α<br />
(<br />
Σ<br />
+ )} , (2.55)<br />
and since α ∈ Ĉ+ 1<br />
implies s α(Σ + ) = s α , we c<strong>on</strong>clude that<br />
w +p<br />
α (λ) = { X ||X| p ∈ D λ s α = s αλ<br />
}<br />
= s(αλ) 1/p . (2.56)<br />
C<strong>on</strong>versely, we have (αλ) 1/p ∈ s (αλ) 1/p = w α<br />
+p<br />
(λ). So<br />
⎛<br />
C + (λ) [ (αλ) 1/p] ∞∑<br />
p<br />
= ⎝<br />
k=n<br />
α k λ k<br />
λ k<br />
⎞<br />
⎠<br />
n≥1<br />
∈ s α , (2.57)<br />
that is, α ∈ Ĉ+ 1 and we have proved (i). We obtain (i)(b) by reas<strong>on</strong>ing as above.<br />
(ii) Assume that αλ ∈ Ĉ 1 .Then<br />
w p α(λ) = { X ||X| p ∈ ∆(λ)s α<br />
}<br />
. (2.58)
1664 BRUNO DE MALAFOSSE<br />
Since ∆(λ) = ∆D λ ,weget∆(λ)s α = ∆s αλ . Now, from αλ ∈ Ĉ 1 we deduce that ∆ is<br />
bijective from s αλ into itself and w p α(λ) = s (αλ) 1/p . C<strong>on</strong>versely, assume that w p α(λ) =<br />
s (αλ) 1/p .Then(αλ) 1/p ∈ s (αλ) 1/p implies that<br />
C(λ)(αλ) ∈ s α , (2.59)<br />
and since D 1/α C(λ)(αλ) ∈ s 1 = l ∞ , we c<strong>on</strong>clude that C(αλ)(αλ) ∈ l ∞ . The proof of<br />
(ii)(b) follows the same lines as in the proof of the necessity in (ii) replacing s αλ by s ◦ αλ .<br />
3. New sets of <strong>sequence</strong>s of the form [A 1 ,A 2 ]. In this secti<strong>on</strong>, we will deal with the<br />
sets<br />
[<br />
A1 (λ),A 2 (µ) ] = { X ∈ s | A 1 (λ) (∣ ∣ A2 (µ)X ∣ ) }<br />
∈ s α , (3.1)<br />
where A 1 and A 2 are of the form C(ξ), C + (ξ), ∆(ξ), or∆ + (ξ) and we give necessary<br />
c<strong>on</strong>diti<strong>on</strong>s to get [A 1 (λ),A 2 (µ)] in the form s γ .<br />
Let λ and µ ∈ U +∗ . For simplificati<strong>on</strong>, we will write throughout this secti<strong>on</strong><br />
[ ] [<br />
A1 ,A 2 = A1 (λ),A 2 (µ) ] = { X ∈ s | A 1 (λ) (∣ ∣ A2 (µ)X ∣ ) }<br />
∈ s α (3.2)<br />
for any matrices<br />
A 1 (λ) ∈ { ∆(λ),∆ + (λ),C(λ),C + (λ) } ,<br />
A 2 (µ) ∈ { ∆(µ),∆ + (µ),C(µ),C + (µ) } .<br />
(3.3)<br />
So we have for instance<br />
[C,∆] = { X ∈ s | C(λ) (∣ ∣) } (<br />
∣ ∆(µ)X ∈ sα = wα (λ) ) ∆(µ) ,.... (3.4)<br />
In all that follows, the c<strong>on</strong>diti<strong>on</strong>s ξ ∈ Γ ,or1/η ∈ Γ for any given <strong>sequence</strong>s ξ and η can<br />
be replaced by the c<strong>on</strong>diti<strong>on</strong>s ξ ∈ Ĉ 1 and η ∈ Ĉ+ 1 .<br />
3.1. Spaces [C,C], [C,∆], [∆,C],and[∆,∆]. For the c<strong>on</strong>venience of the reader we<br />
will write the following identities, where A 1 (λ) and A 2 (µ) are lower triangles and we<br />
will use the c<strong>on</strong>venti<strong>on</strong> µ 0 = 0:<br />
⎧<br />
⎛<br />
⎛ ⎞<br />
⎞<br />
⎫<br />
⎨ ∣ ∣∣∣<br />
[C,C] =<br />
⎩ X ∈ s 1<br />
n∑<br />
⎝<br />
1<br />
m∑<br />
⎬<br />
⎝<br />
λ x k<br />
⎠<br />
⎠<br />
n ∣ µ = αn O(1)<br />
m=1 m ∣<br />
⎭ ,<br />
k=1<br />
⎧<br />
⎛<br />
⎞<br />
⎫<br />
⎨ ∣ ∣∣∣<br />
[C,∆] =<br />
⎩ X ∈ s 1<br />
n∑<br />
∣ ⎬<br />
⎝ ∣ µk x k −µ k−1 x k−1 ⎠ = αn O(1)<br />
λ n<br />
⎭ ,<br />
k=1<br />
⎧<br />
∣ ⎛ ⎞<br />
⎛ ⎞<br />
⎫<br />
⎨ ∣ ∣∣∣∣∣ ∣∣∣<br />
n−1<br />
[∆,C]=<br />
⎩ X ∈ s 1 ∑<br />
−λ n−1<br />
⎝ x i<br />
⎠<br />
µ n−1 ∣ +λ 1<br />
n∑<br />
⎬<br />
⎝<br />
n<br />
x i<br />
⎠<br />
∣ µ n ∣ = α nO(1)<br />
⎭ ,<br />
k=1<br />
[∆,∆] = { X ∈ s |−λ n−1<br />
∣ ∣µn−1<br />
x n−1 −µ n−2 x n−2<br />
∣ ∣+λn<br />
∣ ∣µn x n −µ n−1 x n−1<br />
∣ ∣ = αn O(1) } .<br />
(3.5)<br />
k=1
CALCULATIONS ON SOME SEQUENCE SPACES 1665<br />
Note that for α = e and λ = µ, [C,∆] is the well-known set of <strong>sequence</strong>s that are str<strong>on</strong>gly<br />
bounded, denoted by c ∞ (λ), see[9, 12, 13, 14, 15]. We get the following result.<br />
Theorem 3.1.<br />
(i) If αλ and αλµ ∈ Γ , then<br />
[C,C] = s (αλµ) , (3.6)<br />
(ii) if αλ ∈ Γ , then<br />
[C,∆] = s (α(λ/µ)) , (3.7)<br />
(iii) if α and αµ/λ ∈ Γ , then<br />
[∆,C]= s (α(µ/λ)) , (3.8)<br />
(iv) if α and α/λ ∈ Γ , then<br />
[∆,∆] = s (α(µ/λ)) . (3.9)<br />
Proof.<br />
We have for any given X<br />
C(λ) (∣ ∣ C(µ)X<br />
∣ ∣<br />
)<br />
∈ sα (3.10)<br />
if and <strong>on</strong>ly if C(µ)X ∈ s α<br />
(<br />
C(λ)<br />
)<br />
= s(αλ) ,sinceαλ ∈ Γ .Soweget<br />
X ∈ ∆(µ)s αλ (3.11)<br />
and the c<strong>on</strong>diti<strong>on</strong> αλµ ∈ Γ implies ∆(µ)s αλ = s (αλµ) , which permits us to c<strong>on</strong>clude (i).<br />
(ii) Now, for any given X, the c<strong>on</strong>diti<strong>on</strong> C(λ)(|∆(µ)X|) ∈ s α is equivalent to<br />
∣<br />
∣ ∆(µ)X ∈ ∆(λ)sα = ∆s αλ = s αλ , (3.12)<br />
since αλ ∈ Γ . Thus<br />
X ∈ C(µ)s αλ = D 1/µ Σs αλ = s (α(λ/µ)) . (3.13)<br />
(iii) Similarly, ∆(λ)(|C(µ)X|) ∈ s α if and <strong>on</strong>ly if<br />
∣ ( )<br />
∣ C(µ)X ∈ sα ∆(λ) = C(λ)sα = D 1/λ Σs α = s (α/λ) , (3.14)<br />
since α ∈ Γ .So<br />
X ∈ ∆(µ)s (α/λ) = ∆s (αµ/λ) . (3.15)<br />
We c<strong>on</strong>clude since αµ/λ ∈ Γ implies that ∆s (αµ/λ) = s (αµ/λ) . (iv) Here,<br />
if α ∈ Γ . Thus we have<br />
∆(λ) (∣ ∣ ∆(µ)X<br />
∣ ∣<br />
)<br />
∈ sα if and <strong>on</strong>ly if ∆(µ)X ∈ C(λ)s α = s (α/λ) , (3.16)<br />
since α/λ ∈ Γ . So (iv) holds.<br />
X ∈ C(µ)s (α/λ) = s (α/λµ) (3.17)
1666 BRUNO DE MALAFOSSE<br />
Remark 3.2.<br />
If we define<br />
[<br />
A1 ,A 2<br />
]<br />
0 = { X ∈ s | A 1 (λ) (∣ ∣ A2 (µ)X ∣ ∣ ) ∈ s ◦ α}<br />
, (3.18)<br />
we get the same results as in Theorem 3.1, replacing in each case (i), (ii), (iii), and (iv) s ξ<br />
by s ◦ ξ .<br />
3.2. Sets [∆,∆ + ], [∆,C + ], [C,∆ + ], [∆ + ∆], [∆ + ,C], [∆ + ∆ + ], [C + ,C], [C + ,∆], [C + ,∆ + ],<br />
and [C + ,C + ]. We get immediately from the definiti<strong>on</strong>s of the operators ∆(ξ), ∆ + (η),<br />
C(ξ), and C + (η), the following:<br />
[<br />
∆,∆<br />
+ ] = { ∣ ∣ ∣ ∣<br />
X | λ ∣µn n x n −µ n+1 x ∣−λn−1∣µn−1 n+1 x n−1 −µ n x n = αn O(1) } ,<br />
⎧<br />
∣ ∣ ∣ ⎫<br />
[<br />
∆,C<br />
+ ] ⎨<br />
=<br />
⎩ X | λ ∞∑ ∣∣∣∣∣ ∣∣∣∣∣ x i<br />
∑<br />
∞ ∣∣∣∣∣<br />
x<br />
⎬<br />
i<br />
n<br />
−λ n−1 = α n O(1)<br />
∣ µ<br />
i=n i µ<br />
i=n−1 i<br />
⎭ ,<br />
⎧ ⎛<br />
⎞<br />
⎫<br />
[<br />
C,∆<br />
+ ] ⎨ ∣ ∣∣∣<br />
=<br />
⎩ X 1<br />
n∑<br />
∣ ⎬<br />
⎝ ∣ µk x k −µ k+1 x k+1 ⎠ = αn O(1)<br />
λ n<br />
⎭ ,<br />
k=1<br />
[<br />
∆ + ,∆ ] = { ∣ ∣ ∣ ∣<br />
X | λ ∣µn n x n −µ n−1 x ∣−λn+1∣µn+1 n−1 x n+1 −µ n x n = αn O(1) } ,<br />
⎧ ∣ ∣ ∣ ∣ ⎫<br />
[<br />
∆ + ,C ] ⎨ ∣ ∣∣∣∣∣ ∣∣∣<br />
=<br />
⎩ X λ n n<br />
∑ ∣∣∣∣∣ ∣∣∣∣∣<br />
x i − λ n+1<br />
∑ ∣∣∣∣∣ ⎬<br />
n+1<br />
x i = α n O(1)<br />
µ n µ<br />
i=1<br />
n+1<br />
⎭ ,<br />
i=1<br />
[<br />
∆ + ,∆ +] = { ∣ ∣ ∣ ∣<br />
X | λ ∣µn n x n −µ n+1 x ∣−λn+1∣µn+1 n+1 x n+1 −µ n+2 x n+2 = αn O(1) } ,<br />
⎧ ⎛ ∣ ∣⎞<br />
⎫<br />
[<br />
C + ,C ] ⎨ ∣ ∣∣∣<br />
∞<br />
=<br />
⎩ X ∑ ∣∣∣∣∣<br />
⎝ 1 1 ∑<br />
k ∣∣∣∣∣ ⎬<br />
x i<br />
⎠ = αn O(1)<br />
λ<br />
k=n k µ k<br />
⎭ ,<br />
i=1<br />
⎧<br />
⎫<br />
[<br />
C + ,∆ ] ⎨ ∣ ∣∣∣<br />
∞<br />
=<br />
⎩ X ∑ ( )<br />
1 ∣ ∣ ⎬<br />
∣µk x k −µ k−1 x k−1 = α n O(1)<br />
λ<br />
k=n k<br />
⎭ ,<br />
⎧<br />
⎫<br />
[<br />
C + ,∆ +] ⎨ ∣ ∣∣∣<br />
∞<br />
=<br />
⎩ X ∑ ( )<br />
1 ∣ ∣ ⎬<br />
∣µk x k −µ k+1 x k+1 = α n O(1)<br />
λ<br />
k=n k<br />
⎭ ,<br />
⎧ ⎛ ∣ ∣⎞<br />
⎫<br />
[<br />
C + ,C +] ⎨ ∣ ∣∣∣<br />
∞<br />
=<br />
⎩ X ∑ ∣∣∣∣∣<br />
⎝ 1 ∑<br />
∞ ∣∣∣∣∣<br />
x<br />
⎬<br />
i ⎠ = αn O(1)<br />
λ<br />
k=n k µ<br />
i=k i<br />
⎭ . (3.19)<br />
We can assert the following result, in which we do the c<strong>on</strong>venti<strong>on</strong> α n = 1forn ≤ 0.<br />
Theorem 3.3.<br />
(i) Assume that α ∈ Γ .Then<br />
[<br />
∆,∆<br />
+ ] = s (α/λµ) − if α λµ ∈ Γ ,<br />
[<br />
∆,C<br />
+ ] = s (α(µ/λ)) if λ (3.20)<br />
α ∈ Γ .<br />
(ii) The c<strong>on</strong>diti<strong>on</strong>s αλ ∈ Γ and αλ/µ ∈ Γ together imply<br />
[<br />
C,∆<br />
+ ] = s (α(λ/µ)) −. (3.21)
CALCULATIONS ON SOME SEQUENCE SPACES 1667<br />
(iii) The c<strong>on</strong>diti<strong>on</strong> α/λ ∈ Γ implies<br />
[<br />
∆ + ,∆ ] = s (αn−1 /µ nλ n−1 ) n<br />
= s (1/µ(α/λ) − ). (3.22)<br />
(iv) If α/λ and µ(α/λ) − = (µ n (α n−1 /λ n−1 )) n ∈ Γ , then<br />
[<br />
∆ + ,C ] = s µ(α/λ) −. (3.23)<br />
(v) If α/λ and 1/µ(α/λ) − = (α n−1 /µ n λ n−1 ) n ∈ Γ , then<br />
[<br />
∆ + ,∆ +] = s ((α/λ) − /µ) − = s (α n−2 /λ n−2 µ n−1 ) n<br />
. (3.24)<br />
(vi) If 1/α and αλµ ∈ Γ , then<br />
[<br />
C + ,C ] = s (αλµ) . (3.25)<br />
(vii) If 1/α and αλ ∈ Γ , then<br />
[<br />
C + ,∆ ] = s (α(λ/µ)) . (3.26)<br />
(viii) If 1/α and α(λ/µ) ∈ Γ , then<br />
[<br />
C + ,∆ +] = s (α(λ/µ)) −. (3.27)<br />
(ix) If 1/α and 1/αλ ∈ Γ , then<br />
[<br />
C + ,C +] = s (αλµ) . (3.28)<br />
Proof.<br />
(i) First, for any given X, the c<strong>on</strong>diti<strong>on</strong> ∆(λ)(|∆ + (µ)X|) ∈ s α is equivalent to<br />
∣ ∆ + (µ)X ∣ ( ) ∈ sα ∆(λ) = s(α/λ) , (3.29)<br />
since α ∈ Γ .SoX ∈ s αλ (∆ + (µ)) and applying Corollary 2.9, we c<strong>on</strong>clude the first part<br />
of the proof of (i).<br />
We have ∆(λ)(|C + (µ)X|) ∈ s α if and <strong>on</strong>ly if<br />
∣ C + (µ)X ∣ ∈ C(λ)sα = D 1/λ Σs α . (3.30)<br />
Since α ∈ Γ , we have Σs α = s α and D 1/λ Σs α = s (α/λ) . Then, for α/λ ∈ Γ + , X ∈ [∆,C + ] if<br />
and <strong>on</strong>ly if<br />
X ∈ w +1<br />
(α/λ) (µ) = s (α(µ/λ)). (3.31)<br />
(ii) For any given X, C(λ)(|∆ + (µ)X|) ∈ s α is equivalent to<br />
and since αλ ∈ Γ we have w 1 α(λ) = s αλ .So<br />
if αλ/µ ∈ Γ . Then (ii) is proved.<br />
∆ + (µ)X ∈ w 1 α(λ), (3.32)<br />
X ∈ s αλ<br />
(<br />
∆ + (µ) ) = s (α(λ/µ)) − (3.33)
1668 BRUNO DE MALAFOSSE<br />
(iii) Here, ∆ + (λ)(|∆(µ)X|) ∈ s α if and <strong>on</strong>ly if<br />
∣ (<br />
∣ ∆(µ)X ∈ sα ∆ + (λ) ) = s (α/λ) −, (3.34)<br />
since α/λ ∈ Γ . Thus<br />
X ∈ C(µ)s (α/λ) − = D 1/µ Σs (α/λ) − = s (αn−1 /λ n−1 µ n) (3.35)<br />
if (α/λ) − ∈ Γ , that is, α/λ ∈ Γ .<br />
(iv) If α/λ ∈ Γ ,weget<br />
∆ + (λ) (∣ ∣ C(µ)X<br />
∣ ∣<br />
)<br />
∈ sα ⇐⇒ ∣ ∣ C(µ)X<br />
∣ ∣ ∈ sα<br />
(<br />
∆ + (λ) )<br />
= s (α/λ) − ⇐⇒ X ∈ ∆(µ)s (α/λ) −.<br />
(3.36)<br />
Since µ(α/λ) − ∈ Γ , we c<strong>on</strong>clude that [∆ + ,C]= s (µ(α/λ) − ).<br />
(v) One has<br />
[<br />
∆ + ,∆ +] = { X | ∆ + (<br />
(µ)X ∈ s α ∆ + (λ) )} , (3.37)<br />
and since α/λ ∈ Γ ,weget<br />
We deduce that if α/λ ∈ Γ ,<br />
s α<br />
(<br />
∆ + (λ) ) = s (α/λ) −. (3.38)<br />
[<br />
∆ + ,∆ +] (<br />
= s (α/λ) − ∆ + (µ) ) . (3.39)<br />
Then, from Corollary 2.9, ifα/λ ∈ Γ and (α/λ) − /µ = (α n−1 /λ n−1 µ n ) n ∈ Γ ,<br />
(vi) We have<br />
s (α/λ) −(<br />
∆ + (µ) ) = s ((α/λ) − /µ) − = s (α n−2 /λ n−2 µ n−1 ) n<br />
. (3.40)<br />
C + (λ) (∣ ∣ C(µ)X<br />
∣ ∣<br />
)<br />
∈ sα ⇐⇒ C(µ)X ∈ w +1<br />
α (λ), (3.41)<br />
and since α ∈ Γ + , we have w α<br />
+1 (λ) = s αλ .Thenforαλµ ∈ Γ , X ∈ [C + ,C] if and <strong>on</strong>ly if<br />
(vii) The c<strong>on</strong>diti<strong>on</strong> C + (λ)(|∆(µ)X|) ∈ s α is equivalent to<br />
and since α ∈ Γ + , we have w α<br />
+1 (λ) = s αλ . Thus<br />
since αλ ∈ Γ . So (vii) holds.<br />
X ∈ ∆(µ)s αλ = s (αλµ) . (3.42)<br />
∆(µ)X ∈ w +1<br />
α (λ), (3.43)<br />
X ∈ s αλ<br />
(<br />
∆(µ)<br />
)<br />
= D1/µ Σs αλ = s (α(λ/µ)) , (3.44)
CALCULATIONS ON SOME SEQUENCE SPACES 1669<br />
(viii) First, we have<br />
[<br />
C + ,∆ +] = { X | ∆ + (µ)X ∈ w +1<br />
α (λ) } , (3.45)<br />
and the c<strong>on</strong>diti<strong>on</strong> α ∈ Γ + implies that w α<br />
+1 (λ) = s αλ . Thus<br />
[<br />
C + ,∆ +] = { X | ∆ + } (<br />
(µ)X ∈ s αλ = sαλ ∆ + (µ) ) , (3.46)<br />
and we c<strong>on</strong>clude since<br />
s αλ<br />
(<br />
∆ + (µ) ) = s (αλ/µ) −<br />
for αλ<br />
µ ∈ Γ . (3.47)<br />
(ix) If α ∈ Γ + ,<br />
[<br />
C + ,C +] = { X | C + (µ)X ∈ w α +1 }<br />
(λ) = s αλ = w<br />
+1<br />
αλ<br />
(µ). (3.48)<br />
We c<strong>on</strong>clude that w +1<br />
αλ (µ) = s (αλµ), sinceαλ ∈ Γ + .<br />
Remark 3.4. Note that in Theorem 3.3, we have [A 1 ,A 2 ] = s α (A 1 A 2 ) = (s α (A 1 )) A2<br />
for A 1 ∈{∆(λ),∆ + (λ),C(λ),C + (λ)} and A 2 ∈{∆(µ),∆ + (µ),C(µ),C + (µ)}. For instance,<br />
we have<br />
⎧<br />
⎫<br />
⎨ ∣ ( ∣∣∣<br />
[∆,C]=<br />
⎩ X λn<br />
− λ ) n−1<br />
n−1 ∑<br />
x i + λ ⎬<br />
n<br />
αµ<br />
x n = α n O(1) for<br />
µ n µ n−1 µ<br />
i=1 n<br />
⎭ λ ∈ Γ . (3.49)<br />
Similarly, under the corresp<strong>on</strong>ding c<strong>on</strong>diti<strong>on</strong>s given in Theorems 3.1 and 3.3, weget<br />
[∆,∆] = { ( )<br />
X |−λ n−1 µ n−2 x n−2 +µ n−1 λn +λ n−1 xn−1 −λ n µ n x n = α n O(1) } ,<br />
⎧<br />
⎫<br />
[<br />
∆,C<br />
+ ] ⎨ ∣ ∣∣∣<br />
=<br />
⎩ X λ n<br />
x n + ( )<br />
∞∑ x<br />
⎬<br />
m<br />
λ n −λ n−1 = α n O(1)<br />
µ n µ<br />
m=n−1 m<br />
⎭ ,<br />
[<br />
∆,∆<br />
+ ] = { ( )<br />
X |−λ n−1 µ n−1 x n−1 +µ n λn +λ n−1 xn −λ n µ n+1 x n+1 = α n O(1) } ,<br />
(3.50)<br />
[<br />
∆ + ,∆ ] = { X |−λ n µ n−1 x n−1 + ( λ n +λ n+1<br />
)<br />
µn x n −λ n+1 µ n+1 x n+1 = α n O(1) } .<br />
References<br />
[1] B. de Malafosse, Systèmes linéaires infinis admettant une infinité de soluti<strong>on</strong>s [Infinite linear<br />
systems admitting an infinity of soluti<strong>on</strong>s], Atti Accad. Peloritana Pericolanti Cl. Sci.<br />
Fis. Mat. Natur. 65 (1988), 49–59 (French).<br />
[2] , Some properties of the Cesàro operator in the space s r , Commun. Fac. Sci. Univ.<br />
Ank. Ser. A1 Math. Stat. 48 (1999), no. 1-2, 53–71.<br />
[3] , Bases in <strong>sequence</strong> <strong>spaces</strong> and expansi<strong>on</strong> of a functi<strong>on</strong> in a series of power series,<br />
Mat. Vesnik 52 (2000), no. 3-4, 99–112.<br />
[4] , Applicati<strong>on</strong> of the sum of operators in the commutative case to the infinite matrix<br />
theory, Soochow J. Math. 27 (2001), no. 4, 405–421.<br />
[5] , Properties of <strong>some</strong> sets of <strong>sequence</strong>s and applicati<strong>on</strong> to the <strong>spaces</strong> of bounded difference<br />
<strong>sequence</strong>s of order µ, Hokkaido Math. J. 31 (2002), no. 2, 283–299.<br />
[6] , Recent results in the infinite matrix theory, and applicati<strong>on</strong> to Hill equati<strong>on</strong>,Dem<strong>on</strong>stratio<br />
Math. 35 (2002), no. 1, 11–26.
1670 BRUNO DE MALAFOSSE<br />
[7] , Variati<strong>on</strong> of an element in the matrix of the first difference operator and matrix<br />
transformati<strong>on</strong>s, Novi Sad J. Math. 32 (2002), no. 1, 141–158.<br />
[8] , On <strong>some</strong> BK <strong>spaces</strong>, Int. J. Math. Math. Sci. 2003 (2003), no. 28, 1783–1801.<br />
[9] B. de Malafosse and E. Malkowsky, Sequence <strong>spaces</strong> and inverse of an infinite matrix, Rend.<br />
Circ. Mat. Palermo (2) 51 (2002), no. 2, 277–294.<br />
[10] R. Labbas and B. de Malafosse, On <strong>some</strong> Banach algebra of infinite matrices and applicati<strong>on</strong>s,<br />
Dem<strong>on</strong>stratio Math. 31 (1998), no. 1, 153–168.<br />
[11] I. J. Maddox, Infinite Matrices of Operators, Lecture Notes in Mathematics, vol. 786,<br />
Springer-Verlag, Berlin, 1980.<br />
[12] E. Malkowsky, The c<strong>on</strong>tinuous duals of the <strong>spaces</strong> c 0 (Λ) and c(Λ) for exp<strong>on</strong>entially bounded<br />
<strong>sequence</strong>s Λ, Acta Sci. Math. (Szeged) 61 (1995), no. 1–4, 241–250.<br />
[13] , Linear operators in certain BK <strong>spaces</strong>, Approximati<strong>on</strong> Theory and Functi<strong>on</strong> Series<br />
(Budapest, 1995), Bolyai Soc. Math. Stud., vol. 5, János Bolyai <strong>Mathematical</strong> Society,<br />
Budapest, 1996, pp. 259–273.<br />
[14] F. Móricz, On Λ-str<strong>on</strong>g c<strong>on</strong>vergence of numerical <strong>sequence</strong>s and Fourier series, Acta Math.<br />
Hungar. 54 (1989), no. 3-4, 319–327.<br />
[15] A. Wilansky, Summability through Functi<strong>on</strong>al Analysis, North-Holland Mathematics Studies,<br />
vol. 85, North-Holland Publishing, Amsterdam, 1984.<br />
Bruno de Malafosse: Laboratoire de Mathématiques Appliquées du Havre (LMAH), Université<br />
du Havre, Institut Universitaire de Technologie du Havre, 76610 Le Havre, France<br />
E-mail address: bdemalaf@wanadoo.fr