• Sonuç bulunamadı

A note on vanishing of the functor ext1 for Köthe spaces

N/A
N/A
Protected

Academic year: 2021

Share "A note on vanishing of the functor ext1 for Köthe spaces"

Copied!
7
0
0

Yükleniyor.... (view fulltext now)

Tam metin

(1)

A N o t e o n V a n i s h i n g o f t h e F u n c t o r

E x t I for K 6 t h e S p a c e s

Mefharet Kocatepe

In this note we study the relationship between the vanishing of Ext 1 (A(A), A(A)) and the existence of a regular basis in the Khthe space A(A). We construct an example of a nuclear Khthe space A(A) with no regular basis and such that Extl(A(A), A(A)) = 0. Then we show that for some classes of g 6 t h e spaces A(A), the vanishing of Extl(A(A), A(A)) yields a regular basis for A(A).

I n t r o d u c t i o n

In [7], the K6the spaces A(A) with property (DN) and satisfying the condition Ext I(A(A), A(A)) = 0 have been completely characterized. As a by-product of this characterization it follows that such spaces always have regular bases. It was believed that in this result the condition (DN) was superfluous. In this note we give an example of a nuclear K6the space A(A) such that Extl(A(A), A(A)) = 0, but A(A) has no regular basis. However we show that if A(A) is a direct sum or tensor product of two Dragilev spaces (defined by functions with comparable growth rates) and Extl(A(A), A(A)) = 0, then A(A) has a regular basis. We also show that the same conclusion holds if A(A) is a direct sum or tensor product of a (DN) space and a regular space with property (~).

Notation and terminology

For the terminology not defined here we refer the reader to Dubinsky [4] and Vogt [11].

In the sequel we shall assume that all Khthe spaces A(A) are Fr~chet and Schwartz, i.e. A = (a/k) is such that for all

i,k

6 IN = { 1 , 2 , . . . } we have 0 < a~ < a/k+l and for all k, liml

a~/a~ +1 = O.

The condition (S;) (called (S*) in [8]) was defined in [11] by Vogt. Let A(A) and A(B) be two Khthe spaces. The pair (A(A), A(B)) is said to satisfy the condition (S;) (briefly (A(A), A(B)) 6 (S~)) if the following holds:

n~ /

(2)

K O C A T E P E

It is easy to see that this condition is equivalent to the following a K ( a ~ a ~ ) V u 3 k V I f 3 n 3 i 0 , j 0 V i > i 0 , V j > / 0 : ~ < m a x \ b ~ : , b] "

It was shown in [11] and [8] that (~(A), ~(B)) satisfies (S;) if and only if E x t l ( ~ ( A ) , ~(B)) = 0 if and only if every short exact sequence 0 --+ ~ ( S ) F ---* )~(A) ---* 0 of Fr~chet spaces splits.

The conditions (DN) and (~) were defined by Vogt [10] and Wagner [12] to characterize subspaces and quotients of nuclear and stable power series spaces of infinite type and finite type respectively. Here we state KSthe space versions of these conditions which in this case turn out to be the same as the conditions (dl) and (d~) of Dragilev [3]. A g S t h e space ~(A) is said to have the property (DN) if 3n0 Vm 3n, C > 0 Vi: (a[n) 2 < Ca~~

( ~ ) i f V p 3 q V k 3 C > 0 V i : C(a~) ~> k p a i a i .

A Khthe space )~(A) is said to be regular if for all i and k, a~+l/a~ <

a k + l / . k i+l I t * i + l and it is called pseudo regular if

ai <M.EL. Vp 3q Vr > q 3s > p 3 M > O : i <- J =~ aq -- af

Regularity was defined in [3] and pseudo regularity in [2]. It is obvious that pseudo regularity is weaker than regularity (it is not known whether it is strictly weaker.)

In [2], it was shown that if ~ = ~(A) is a nuclear, pseudo regular Khthe space, then 6(,~) = $ 9 Sx. Here 6()~) denotes the diametral dimension of (see [1]) and Sx is the Khthe dual (or a-dual) of ,~. It is well-known that if )~ is nuclear, then ~f($) C 6(s) where s denotes the space of rapidly decreasing sequences, i.e. s = )~(B) where b~ = e x p k l o g i and 6(s) = s. Hence for a nuclear, pseudo regular Khthe space ~ = $(A) we have $ 9 Sx C s and this last condition is equivalent to the following:

a k el log i

V k V s s u p - - < c o .

(See proof of (3.3) and (5.1) (i) in [9].)

L ] ( a , r)-spaces (also called Dragilev spaces) were introduced in [3]. Let

f

: fi --. R be an odd, strictly increasing, logarithmically convex function (i.e. log o f o exp is convex), a = (ai) be an exponent sequence, i.e. 0 < ai / z oo, rk z z r where r = - 1 , 0 , 1 or oo. Then L ! ( a , r ) is defined as the KSthe space

A(A) generated by the matrix a~ = exp f(rkai). If f is logarithmically convex,

then it satisfies one of the following conditions:

Va > 1, lin~-..oo f ( a z ) / f ( z ) = or In this case f is called rapidly increasing,

Va > 1, lirn=-.oo f ( a x ) / f ( z ) < o0. In this case f is called slowly increasing.

When f is slowly increasing L ! (a, r) is isomorphic to a power series space and

(3)

If f and g are two functions as in the definition of Dragilev spaces, we write f >- g or g -~ f if g - i f is a rapidly increasing function,

f ~ g if g - i f and f - l g are b o t h logarithmically convex and slowly increasing. T h e e x a m p l e

First we construct a nuclear Khthe space )~(A) with no pseudo regular basis and with the p r o p e r t y t h a t E x t l ( A ( A ) , ~(A)) = 0.

P r o p o s i t i o n 1. Let (~i) be any sequence such that 0 < ~i /2 oo and the set of c o

finite limit points of the set {c~i/o~j : i , j E N} is bounded. Let N = Uu=lNu be a partition of N into pairwise disjoint infinite sets N~. Let rk / 2 oo and a~ "~ 0. For each i E N, let u = u(i) be the unique index such that i E N~. Let

p~ = { rk i f u < k - ~ if k < u and set a ik = exp pikcq. Then

Extl(k,(A), ),(A))

= 0.

Proof. E x t l ( A ( A ) , A(A)) = 0 if and only if VI~ 3k V K 3n 3io,jo :Vi > i0,

vj >j0

either (I): c~_j_j < p~ - pK or (II): pK _ p~ < a_zj

. , - p ; - - - . ,

holds.

Let 0 < a _< A < o o be such t h a t if a is a n y finite non-zero limit point of the set { a i / ~ j : i , j E N} then a _< c~ < A.

f f k o" k l < a . rk +cr~' a ~ a k

]

rn -- rK rn -- rK rn + ffK rn + CrK rn + a K ~ rK + f f k ' •k - - f f K ' rK - - r k ~ rK +r f f k - - - ~ K ) Given #, we choose k so t h a t m a x - - , r k 7"# Given K , we choose n so t h a t A < m i n (r_pn - r___K, \ rK -- rk

Given i and j , assume i E Nu and j E Nv. We have several possibilities for the positions of u a n d v. If u _< k, v < #, then

(I): ~__AJ < r , - r______gg, (II): r / < - r_____A~ _ <

~j_

Oli rK - - rk rk - - r ~ o q

In general if u < n and v is a r b i t r a r y we have ,,(I~: ~" _< r , . . . .

O q . . .

and by o u r choice of n for all values of i and j for which a j / a i a c c u m u l a t e a b o u t a finite limit point, (I) holds. For the others (II) holds.

(4)

K O C A T E P E

I f n < u, t h e n (II) becomes (a~ - a K ) / ( " ") < a j / a i and by our choice of k, for all values o f i and j for which a j / c q a c c u m u l a t e a b o u t a limit point which is bigger t h a n or equal to a, (II) holds. For the other values of i and j , (I) holds. []

E x a m p l e . Let (in) be a subsequence of N such t h a t lirnn-..~ log in+i~ log in = c~. I f ( i , ) is such a sequence t h e n for all large n we have i, >_ n n. Let i0 = 0 and define

e e i = c q . = l o g i n for i n - l + l < i < i , , n = 1 , 2 , . . . .

Let

N i = { i : i # i n , n = l , 2 , . . . } , N ' l = { i n : n = l , 2 , . . . } a n d p a r t i t i o n l~t~ into pairwise disjoint infinite sets NI2 U NI3 U . . .

Since the set { ~ J % ) has only three limit points (r, a m e l y 0, 1 and oo), the K S t h e space A(A) defined as in P r o p o s i t i o n 1 has the p r o p e r t y t h a t E x t l ( ) ~ ( a ) , A ( a ) ) = 0. Now we show t h a t A(A) is nuclear. For this observe the following:

(i) Vc > O, E n % l ~ v z . < ~ . = 1 1 oo n-- ~ < ~ .

O0 ~,, I O0 tn -- ~n--I <

(ii) Vc > 1,

Ei=I

~ e l~a = E , = i E i = i . _ , + i e c o t i - E , = i - - e~=,.

o o

G i v e n k we find g such t h a t rl > rk + 1. I f i E N,, we have

1 i f u < k

a k. e (rt-rk)~

--, _ 1 i f k < u < g

a-~i - - e(rL+ak )c,i

1 if g < U. e(ak-a~)ai T h e n for u > 2 we have a n d so a l < 1 a~ - e ( ~ - ~ , ) ~ , ~ a ~ < ~---, 1 c~ 1 a . ,

a _ . , i = l a~ -- e(r'-r~) o' + y~" e(~k-~,)~,. < ~ '

i E N I n = l

Finally we show t h a t A(A) does not have a pseudo regular basis. I f it did we would have

a k e g log i

V k V ~ 3 r n : s u p - - - C < ~ . a~ n

Let k = g = 1. Given any m > k, let u > m. T h e n

a/ke g log i

sup - - -- sup e t l ~ = sup e ( g - ( a k - a ' ~ ) ) l ~ = CO.

(5)

S o m e o b s e r v a t i o n s

Next we indicate some cases in which the vanishing of Exti(A(A), A(A)) yields a regular basis for )~(A).

T h e o r e m 1.

Let A(A) and A(B) be Schwartz regular Kd'the spaces with A(A)

having property

(DN)

and )~(B) having property (-~). If

Extl(A(B), A(A)) = 0,

then A(A) @ )~(B) and ) ~ ( A ) ~ A ( B ) have regular bases,

Proof.

After a suitable standardization (see [8]) we may assume that

b]

(b~ -1

b~+l~

a-~ < max ~a~_l, ai ~ ] (1)

k - i ~ k + i i i (2)

(a~) 2 ~_ a i

ui

, al =

> _ - : _j , b ~ = l (3)

for all k, i and j .

Let (e~) (resp. (f/)) be the natural basis vectors in ,X(A) (resp. ,~(B)). The rearrangement of the sequence of vectors el, fl,e2, f 2 , . . , corresponding to a al, bl, a2, b2,.., is easily seen to nondecreasing rearrangement of the sequence 2 2 2 2

produce a regular basis of A(A) @ A(B).

Next we consider

A(A)~,~)~(B)

which is isomorphic to ,X(C) where c/~j = k k We let

a i bj.

: = n X n \ z .

I = {(i, j ) : a i _ For (i, j ) E I by (1) we have

~i__t__

< l~__

a~ - bj

hence because of (2) and (3)

_ l + l

hk+l

~'i.._L_ < L~_j~. and s o for all k for all t, k

(4)

a~ _< b~ for all k and therefore

b] ~_ a~b] <_ b~

.k.

For ( i , j ) E J, because of (2) and (3) we have the reverse inequality for (4), hence

a i a i b j ~ a i o i

So A(C) ~_ ,~(C1) (9 ,~(C2) where A(Cx) has property (DN), ,~(C2) has property (~) and Extl()~(C2),)~(Ci)) = 0. This leads us to the first case. n

(6)

K O C A T E P E

Since for a K6the space $(A) with (DN) from Extl(A(A), ,~(A)) = 0 it follows t h a t ~(A) has a regular basis (see [7]), we immediately obtain the following Corollary.

C o r o l l a r y . Let )~(A) and A(B) be Schwartz g6"the spaces with properties (DN) and (-~) respectively. Assume Extl()~(A) @ A(B), ~(A) (9 ~(B)) = 0 and )~(B) has a regular basis. Then )~(A) @ A(B) and A ( A ) ~ , A ( B ) have regular bases.

If E = n / ( a , r) and F = Lg(b, s) are two Dragilev spaces such that f -~ g or

f ~- g or f ~ g, then complete characterizations for ExtZ(E, F ) = 0 have been given by Hebbecker [5] (see also [6]). By using these characterizations and a proof similar to t h a t of T h e o r e m 1 we obtain the following.

T h e o r e m 2. Let E = L f ( a , r ) and F = ng(b,s ) be two Dragilev spaces such that either f ~ g or f ~- g or f ~ g and ExtZ(E @ F , E (g F) = O. Then E (g F and E ~ , F have regular bases.

We wish to thank the referee for simplifying the proof of T h e o r e m 1.

R e f e r e n c e s

[1] Bessaga, C.: Some remarks on Dragilev's theorem. Studia Math. 31, 307-318 (1968)

[2] Crone, L., Dubinsky, E., Robinson, W.B.: Regular bases in products of power series spaces. J. Funct. Anal. 24, 211-222 (1977)

[3] Dragilev, M.M.: On regular bases in nuclear spaces. Math. Sb. 68,153-173 (1965) (Amer. Math. Soc. Transl. 93, 61-82 (1970))

[4] Dubinsky, E.: The structure of nuclear Frdehet spaces. Lecture Notes in Mathe- matics 720 (1979)

[5] Hebbecker, J.: Auswertung der Splittingbedingungen (S[) und (S~) ffir Poten- zreihenr~ume und LI-Riume. Diplomarbeit, WuppertaJ, 1984

[6] Kocatepe, M., Nurlu, Z.: Some special K6the spaces. Advances in the theory of Fr~chet spaces (ed: T. Terzio~lu) 269-296, NATO ASI Series, Series C 287 (1989) [7] Krone, J.: Zur topologischen Charakterisierung yon Unter- und Quotien- tenr~umen spezieller nuklearer Kbtheriume mit der Splittingmethode. Diplo- marbeit, Wuppertal, 1984

[8] Krone, J., Vogt, D.: The splitting relation for Kbthe spaces. Math. Z. 180,387- 400 (1985)

[9] Robinson, W.B.: Relationships between ~-nuclearity and pseudo-/t-nuclearity. Trans. Amer. Math. Soc. 201,291-303 (1975)

(7)

[11] Vogt, D.: On the functors Ext:(E, F) for Fr6chet spaces. Studia Math. 85, 163-

197

(1987)

[12] Wagner, M.J.: Quotientenr~ume von stab[len Potenzreihenr~umen endlichen Typs. manus, math. 31, 97-109 (1980)

Mefharet Kocatepe Bilkent University

Department of Mathematics Faculty of Engineering and Science 06533 Bilkent, Ankara, T U R K E Y

(Received June 26, 1990;

Referanslar

Benzer Belgeler

Also, the person named as Bairambec (by the traveller) whom Zahid Bey fought against was Dulkadiroğlu Behram Bey, who was the deputy of Kurd Bey, the commander of Shah

Bütün bu açıklamalardan sonra İbn Sînâ, Zatıyla Zorunlu Varlık’ın varlığa gelmesi, yani Zatının taayyünü bakımından bir (vahid) olduğu ve hiçbir yönden çoklukla

In this study, 201 thermophilic bacteria that were isolated from natural hot springs in and around Aydin and registered in Adnan Menderes University Department of Biology

The boron applications caused the available boron contents of soil reach to the toxic level and boron toxicity symptoms were appeared on the plant.. Effect of boron

Overall, the results on political factors support the hypothesis that political constraints (parliamentary democracies and systems with a large number of veto players) in

Here we extend this result (under some additional restriction to p and q) to the case when the …rst space is isomorphic to a complemented subspace of the second one.. First we prove

Zahariuta, Bounded operators and isomorphisms of Cartesian products of Köthe spaces, C.R.. Dragilev, On regular bases in nuclear

“Sınıf Öğretmeni Adaylarının Kişisel Ve Aile Özellikleri İle Öğrenim Gördükleri Program, Öğretmenlik Mesleği Ve Yaşama İlişkin Görüşleri / Personal