Internat. J. Math. & Math. Sci. VOL. 13 NO. 2 (1990) 243-246 A NOTE ON SOME SPACES 243 L OF DISTRIBUTIONS WITH LAPLACE TRANSFORM SALVADOR PEREZ ESTEVA Instituto de Matemticas Universidad Nacional Mxico, D.F. Autnoma de Mxico 04510 M6xico (Received March 2, 1989) In this paper we calculate the dual of the spaces of distributions ABSTRACT. [I ]. introduced in L Then we prove that KEY WORDS AND PHRASES. is the dual of a subspace of D’ and denote by a L spaces Laoy -Y PY S’ and L be the classical Schwartz’s L2OR), then we define Laoy Dp Since La PY were introduced as follows: L is the subspace of f spaces of distributions in In (Prez-Esteva [I ]) the Laplace transformation. where e [a DY py a of functions f with supp f C [a,) and a is a Hilbert space with the inner product OR) loc -Tx. L oy -y (x)=e I (f,g) on Primary 44A35, Secondary 44AI0 INTRODUCTION Let e C OR). Convolution, Laplace Transform, Strict Inductive Limit. 1980 MATHEMATICS SUBJECT CLASSIFICATION CODE. I. L e_2 fg dx where Dp is the distributional derivative of order p. oY a is bijective, we can copy the Hilbert space structure of L oY DP[a We have the continuous inclusions La C Lb for a>b C a L qy, if p pY L Hence for p {0,I,...} a PY <q the strict inductive limit L makes sense. py’ ind lim a PY La PY Then L Y ind lim p L ind lim PY p L-p PY is also well defined. In[l] it was studied the spaces of distributions f is continuous. f,g: [ L Y g for which the convolution S.P. ESTEVA 244 of C OR). Sy will denote the norm of N (Sy)’= and conclude that L2OR), y Ly, S 7 which will be will be the set of nonegative integers. [ THE DUAL OF DEFINITION I. such that which turns out to be a subspace Y Then we prove the reflexivity of ll’II is the main result of the paper, 2 assumed to be a positive constant, and 2. L the strong dual of Here we describe Let a,)e_yg he the space of all complex measurable functions L E 20R characteristic function of for every E a [a,). We provide pa(g lla,)e_yg where g in stands for the a,) with the topology given by the L seminorms S Next we denote by n N. S and n g LEMMA I. Ly and Sy are Frechet Sy, we have that Sy C L’, Let {gp}p 6N Hence gp+l by go a 6 Fix Dnf L Y for every , n E N p E N, e-2y f gpdx, Dng there exists f 6 L llocR) gp L for any such that L o satisfies is determined PROOF. a 2 spaces and since then for every (DPf) The sequence such that Y by the system of seminorms II a,)e_y Dngll pan(g It is clear that L the subspace of Define the topology of m. a and -Dgp + 2ygp 6 Sy. p 6 N p 6 N. 6 Then (2.1) (Ly)’ and there exists gpa 6 La such that (DPf) If a < b, we have [ e-2yf gpa b L C pY (DPf) for DPf 6 Lbpy’ which a L py’ then I e_2yf gpb I dx gpa is the restriction of belongs to Let L 6 PY e 2yf X b ,) X[ b,=o)gpa gpa a, =), to gp then and (DPf) D. Since gpa dx shows that gpb If 6 ia DPf dx, J e_2yf gpdx DP+I 6 Lp+ Y ] L PY DPf E L py we have Da gpa is well defined, SPACES OF DISTRIBUTIONS WITH LAPLACE TRANSFORM i e_2ygp+l I e_2yDgp (DP+I) dx ]{D(e_2y) <-e_2{Dgp <-,’> where + 2y e_2 Y}gp dx dx > + 2y e_2{g p D represents the duality between gp+l 245 and D’. It follows that -Dgp + 2ygp or -D(e-2y gp) e-2y gp+l gp Hence, every LEMMA 2. H + 2yl. -D g 6 S Let is well defined in PROOF. {f and Y H be the differential operator defined by Then the functional I (DPf) sequence S. belongs to Let n n6 N L f 6 and is continuous. Y f 6 C p H(P)gdx’ e-2y f Laoy be such that converging to f f Dh b h 6 with L oy if in b Loy. There exists a < a. Let n (x) b ioy’ n uous, we have that Then D’n f and h) f {n}n6N n fn dy and since the inclusion f h converges to in f e_2yh H(g)dx Lim I e_2yf g dx Limf e_27fngdx f e_2yn Lo. Lb o C Lb is contin- If follows that H(g) dx (2.2) (2.3) On the other hand e_2.(on n D(e-2g) tt(g)dx -n(B)e-2y(B)g(B) + fn e-2yg (2.4) dx b But we have the estimate ]g(x) < Ig(b)] + 1/2 ey(x) ]lb,oo)e_y(Dg Yg)](x-b) Hence e_2yn dx H(g)dx e_2yfn g dx From (2.2) and (2.3) it follows that f e_2yfgdx I e_2,h H(g)dx for x >b S.P. ESTEVA 246 By induction we obtain ; if DPh f f e_2yh H(P)(g) e_2yfgdx f,h E and L oy DPf Finally, if (2.6) dx Dqh f,h E L with and q p, f then Dq-Ph, hence by (2.6) we have e_2yf H (P) (g)dx e_2yh H (q) (g)dx is well defined and it is clearly continuous. Thus THEOREM I. PROOF. _ is and 2 we know that By lemmas L, Ly) strong topology Ly The strong dual of Sy. Ly’ Sy. It remains to prove that the coincides with the topology T of First notice that Sy. is defined by the system of seminorms I x[ ,)_ "(P) (g)ll 2’ Let V {g E Sy: qap(g) qap (g) a l a Fix in L oy’ and p then the set DPu II B CE Hence {g p DPu Sy: q_p g B , I}. pY N p l, U Denote by L and hence in the unit ball If g B U we have f I<DPf, (p) (g a,m) H )112 < c B(L’,, Ly). C V Ly. Now, let B be a bounded set in and i -p and is bounded there (see Kucera, McKennon 2 ). B C L PY for some g >0 where U is the unit ball in L -p. Let oY V implies for f E EU that then g (g) < P B It follows that Then for some L is bounded in I e-2yf H(P)(g)dxl Thus Then B B), then for every (the polar of V N. a N, E-I}, <DPf, g>= so we proved that COROLLARY I. L . If e_2yf li(p)(g)dx] < V C B This completes the proof. is the strong dual of S Y Y PROOF. By (Kucera, McKennon 2 ], Theorem 4) Hence the corollary follows from Theorem I. we know that L is reflexive. REFERENCES S. Prez-Esteva, "Convolution Operators for the one-sided Laplace TransformatioH’ Casopis pro pestovn matematiky, Vol. 110(1985), 69-76. [2 J. Kucera, K. McKennon, "Dieudonn-Scwartz theorem on bounded sets in inductive limits", Proc. Amer, Math. Soc. 78(1980), 366-368. [3 [4 J. Kucera, "Multiple Laplace Integral", Czech. Math. J. 19(1969), 181-189. L. Schwartz, Theory of Distributions, Hermann, Paris, 1966.