Internat. J. Math. & Math. Sci. (1985) 151-163 151 Vol. 8 No. BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL REFLECTION OF THE ARGUMENT EQUATIONS WITH JOSEPH WIENER and A.R. AFTABIZADEH Department of Mathematics Pan American University Edinburg, Texas 78539 (Received July ABSTRACT. 30, 1984) Linear and nonlinear boundary value problems for differential equations with reflection of the argument are considered. KEY WORDS AND PHRASES. Functional Differential Equation, Boundn. Value Problems, Exis fence, Uniqueness. 1980 MATHEMATICS SUBJECT CLASSIFICATIOi CODE. I. 34KI0. INTRODUCTION. In [1-4] a method has been discovered for the study of functional differential equations whose argument deviations are involutions. Important in their own right, they have applications in the investigation of stability of differential-difference equations. Differential equations with involutions can be transformed by differen- tiation to higher order ordinary differential equations data initial or boundary conditions. been studied in numerous papers. and, hence, admit of point Initial value problems for such equations have However, boundary value problems even for differ- ential equations with reflection of the argument have not been considered yet. The purpose of this paper is to discuss existence and uniqueness of solutions of y" where f e C[[-a, a]xRxR, R], a > f(x, y(x), y(-x)), (I.I) O, with the following types of boundary conditions y(-a) YO" y(a) (1.2) Yl or y’(-a) where h, k O, h + k O. hy(-a) 0, y’(a) + ky(a) 0, (1.3) 152 2. J. WIENER AND A. R. AFTABIZADEH PRELIMINARY RESULTS. First, we prove a sequence of lemmas for the linear case, a(x)y(x) + b(x)y(-x) + c(x), y"(x) which are needed in order to prove our results for the general equations of the form (I.). Before we proceed further, we present some results without proof, which help to simplify the proofs of our results. LEMMA 2.1. [5, pp.182]. fl If y(O) i 2 y (x) dx < 0 LEMMA 2.2. , [6]. f(t, x, y) If [a, b]xP with respect to x and y on (t, x, y),(t, ) 1 C I[0, i], then 2dx. [y’ (x) 0 is continuous and has continuous first partials where P is an open convex set, then for [a, b]xP, f(t,,) f(t,x,y) 0 and y(x) y(1) s(t))(x- ) + f2(t,r(t), f3(t,(t),(t))(y- y), where f2(t,r(t),s(t)) r), y + x + (1 f2(t, 0 (I and fl f3(t,(t),(t)) are continuous functions on between x and x, and 0 LEMMA 2.3. _< Y x + (I f3(t, [a, b]xP ), y + ))dT, (I s(t), (t) between y and with , <_ I. The homogeneous boundary value problem 0, U together with u(a) u(b) O, has the Green’s function G(x, t), defined by [(b-x)(t-a), a <_ t < x < b (b-t)(x-a), a <_ x 1 b-a C(x, t) <_ t and the following estimates hold true: [G(x a t) [dt < (b-a)2 8 b a IGx(X t)[dt < b-a 2 Now consider the second order linear differential equation _< b, r(t), (t) BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL EQUATIONS y" 153 (2.1) a(x)y + b(x) with where a(x), b(x) e C[[a, b], R]. [F]. LEMMA 2.4. O, h, k > O, h + k > 0, 0, y’(b) + ky(b) hy(a) y’(a) Then [a, b], Suppose a(x) _> m > 0 on then boundary value problem (2.1) has a unique solution satisfying < i suply(x) 3. suplb(x) m x e [a b]. MAIN RESULTS. We are now in a position to state our results. LEMMA 3.1. b y2(x)dx cl[a, b], then [Y ’(x)]2dx" a I This follows easily from Lemma 2.1. LEMMA 3.2. Since fb <-- [J y(b) y(a) cl[a, 0 and y e y(b) If y(a) suply(x) PROOF. -- Ib (b_a)2 < a PROOF. O, and y(x) e y(b) If y(a) b], then [Y’(X)]2dx] I/2 a <_ x <_ b. a 0, then one has x iu y’(t)dt 2y(x) la y’(t)dt, x or 21y(x) <- I lY’(t) Idt lY’(t) Idt + ly’(t) a x a or sup lY(X) ly’ (t)ldt, <_ - a < x _< b. a Using the Cauchy-Schwartz inequality we have suply(x) <1/2[ Thus suply(x) <-LEMMA 3.3. ib dt a ]i/ 2 ib [y, (t)] 2d t ]1/2 a <_x <_b a b I [y’(x)]2dx]1/2 a<x<b. a Consider equation (2.1) with the boundary condition y(a) y(b) O, (3.1) 154 J. WIENER AND A. R. AFTABIZADEH (a)2 >_- a 0 > where a(x) suply(x) (b-a)2 <-2 y(x) of (2.1) and (3.1) satisfies Then any solution suplb(x) I, [2-ao(b-a) 2] a <_ x <_ (3.2) b. On multiplying (2.1) by y(x), and integrating the result from a to b, we PROOF. find, because of (3.1), [y’(x)]2dx b(x)y(x)dx, a(x)y (x)dx a a a or a [y’(x)]2dx <_ a0 IY (x) (x)dx + suplb(x) dx" a a Applying Lemma 3.1 and the Cauchy-Schwartz inequality, we get [Y’(X)]2dx --< a (b-a) 0 n a 2 ib 2 [Y’(X)]2dx (b-a)3/2 + a sup lbCx)l I 1/2 b [y’Cx)] 2 dx] a or [y’ (x) 2 dx ]I / n(b-a/2 2 < 2 a -ao (b-a) 2 suplbCx) I, a<x<_b. Lemma 3.2 and the above inequality then imply inequality (3.2), Suppose a(x) and b(x) satisfy all conditions of Lemma 3.3. LEMMA 3.4. Then problem (2.1) with the boundary conditions y(a) y(b) YI’ (3.2) Y2 has a unique solution. PROOF. First, we show the uniqueness. (2.1), (3.2). Let R(x) u(x) R"(x) By Lemma 3.3, R(x) unique solution. Suppose u(x) and v(x) are solutions of v(x), then a(x)R(x), O, which implies u(x) R(a) R(b) v(x). So problem (2.1), (3.2) has a O. To prove the existence, let u(x) and v(x) be solutions of the following initial value problems (i) u"(x) a(x)u(x) + b(x), u(a) YI’ u’(a) 0; (ii) v"(x) a(x)v(x) v(a) O, v’(a) I. We notice that u(x) and v(x) exist and are unique. v(b) v’(a) O, then from v(a) i. Moreover, v(b) O, (ii) and Lemma 3.3 we have v(x) Therefore, by linearity, y(x) u(x) + Y2-u(b) v(b) v(x) O, because if 0 which contradicts BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL EQUATIONS defines the solution of the problem (2.1), (3.2). 155 Proof is complete. Let us now, consider the second order linear functional differential equation C[-a, a], where a(x), b(x), c(x) tion on [-a, a] 2. < O. We shall show that, under certain condia boundary condition has a unique solu- and obtain an estimate for such solution. Ib(x) Suppose that a(x) >-m, LEMMA 3.5. 4a2(m+n) a > b(x), equation (3.3) with tions on a(x) and (3.3) a(x)y(x) + b(x)y(-x) + c(x) y"(x) < n, b(x) # O, x e [-a, a], and Then any solution of equation (3.3) with the boundary conditions y(a) y(-a) (3.4) O, satisfies suply(x) 2a 2 < 2_4a 2 suplc(x) (3.5) -a < x < a. (m+n) On multiplying (3.3) by y(x) and integrating the result from -a to a we PROOF. have [y’(x)]2dx a(x) -a (x)dx b(x)y(x)y(-x)dx -a a c(x)y(x)dx, a or [y’ (x) ]2dx < lY(X) lY(-X) Idx + suplc(x)l (x)dx + n m a -a -a ly(x) -a Now, using Lemma 3.1, the Cauchy-Schwartz inequality, and the facts that < 1 [y 2 (x) ly(x) IIy(-x)l + y2 (-x)], and 2 y (x)dx -a I_ a y2(-x)dx, a we obtain a [| [y’(x)]2dx] I/2 _< J_ a n(2a)3/2 2-4a2(m+n) sup[c(x) I, -a <_ x<a. (3.6) Applying Lemma (3.2) and inequality (3.6) we get inequality (3.5). Having Lemma 3.5, we can prove the following theorem. THEOREM 3.1. In addition to the assumptions of Lemma 3.5, suppose that a(x) and b(x) are even functions on [-a, a]. y(-a) has a unique solution. Then equation (3.3) with the boundary conditions YI’ y(a) Y2 (3.7) J. WIENER AND A. R. AFTABIZADEH 156 PROOF. Uniqueness follows from the fact that if u(x) and v(x) are two solutions of (3.3), (3.7), then R(x) v(x) implies u(x) R"(x) a(x)R(x) + b(x)R(-x), R-a) R(a) Hence, from Lemma 3.5, R(x) 0. v(x). 0 and u(x) Now, we show that problem (3.3), (3.7) in fact has a solution. (3.8) y(-x). y(x) u(x) Let Then (3.9) y"(-x). y"(x) u"(x) From (3.3) and (3.9) we have a(x)y(x) + b(x)y(-x) u"(x) Since a(x) and b(x) are c(-x). even, then a(x)[y(x) u"(x) a(-x)y(-x)-b(-x)y(x) + c(x) y(-x)] b(x)[y(x) y(-x)] + c(x) c(-x), or by (3.8), u"(x) [a(x) u(-a) Yl b(x)]u(x) + c(x) c(-x), (3.10) and Problem (3.10), (3.11) solution u(x). Y2’ u(a) is a form of Y2 (3.11) Yl (2.1), (3.2), then by Lemma 3.4, it has a unique Hence y(-x) is given by y(-x) y(x) u(x). This implies that y"(x) [a(x) + b(x)]y(x) + c(x) y(-a) YI’ b(x)u(x) (3.12) and Again by Lemma 3.4, Problem tion of (3.3), (3.7). y(a) (3.13) Y2" (3.12), (3.13) has a unique solution which is the solu- Proof is complete. Now consider the following second order linear functional differential equation y"(x) where a(x) # 0 on [-a, a]. a(x)y(-x) + b(x) (3.14) By differentiation and algebraic elimination this equa- tion can be reduced to the fourth order differential equation y where (4) (x) A(x)y"’(x) + B(x)y"(x) + C(x)y(x) + D(x), x 6 I-a, a], (3.15) BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL EQUATIONS A(x) 2a’ (x) a(x) B(x) (x)l ra’ ---, C(x) a(x)a(-x), a’(x) 2 a(x) O(x) -a(x)[b(-x) + l " 1 a ’(x) %- 157 a2_x_ () , rb(x) La- By a solution of (3.14) we mean a solution that is four times dlfferentiable. We shall show that equation (3.15) with the boundary conditions AI, y(-a) y(a) A2, y"(-a) Sl, y"(a) B (3.16) 2 We use the method given by R. A. Usmanl [8]. has a unique solution. First we need the following Lemma. LEMMA 3.6. Consider the fourth order linear differential equation (3.15) with the boundary condltion y(.-a) where A(x) e cl[-a, Ic(x)l If _< n. y"(-a) y(a) y"(a) O, (3.17) a], B(x), C(x), D(x) e C[-a, a], + 16na 4 < 4m2a 2 then any solution of 1 A’(x) l(x) <_m and 4 (3.18) (3.15), (3.17) satisfies 4 suply(x) <PROOF. 4 -4m2a2-16na 4 suplD(x) l, -a <_x_< a. (3.19) Let yl! (3.20) Z. On multiplying (3,20) by y(x) and integrating the result from -a to a, we have [y (x) 2dx yzdx, a or _a [y’(x)]2dx <[ a i_a y2(x)dx ]I/2 i_a a 2 z (x)dx ]1/2 a Applying Lemma 3.1, we obtain [y,(x)]Zdx a Also, from (3,20) and (3.15) 2 _< n2 [z,(x)]2dx a 2 (3.21) J. WIENER AND A. R. AFTABIZADEH 158 z" A(x)z’ + B(x)z + C(x)y + D(x), z(a) z(-a) (3.22) O. In a similar manner, 3 [z’ (x) ]2dx suplD(x) 4_ 2-16ha 4_4ma a a <_ x <_ a. (3.23) From Lemma 3.2, (3.21) and (3.23) inequality (3.19) follows. THEOREM 3.2. (3.15), (3.16) has a unique solution. (3.15) and (3.16). (4) Then it is easily seen that (x) A(x)"’ (x)+B(x)"(x)+C(x)(x), (-a) (x) Now, from Lemma 3.6 and (3.24) and u(x) Then problem Assume that there exist two distinct functions u(x) and v(x) satisfying PROOF. 4 Suppose all assumptions of Lemma 3.6 hold true. v(x) on [-a, a]. it follows that u(x) -v(x) satisfies (a) "(-a) supl(x) <_ O, This shows that problem "(a) which proves (3.15), (3.16) has O. (3.24) (x) 0 at most one solution. In order to prove that (3.15), (3.16) indeed has a solution, we define functions Yi(X), 4, as solutions of the respective initial value problems: i, i (1) yI(4) A(x) Y’ +B (x) Y’+C(x) Yl+D(x) (ii) y (ill) y4) =A(x)y’ +B (x) y+C (x) y3, (iv) Y4 =A(x) (4) +B (x) Y2 Yl (-a) =AI YI’ (-a) yl (-a)Y[ (-a) 0 Y2’(-a)=l’ y2(-a)-Y2"(-a)fy (-a)=0; Y(-a) B 1 Y3(-a)ffiy (-a)y’ (-a) O A(x)"’+B(x)y2+C(x)y Y4 4 -Y4"’ (-a) 1 Y4 (-a)Y (-a)--y4’’ (-a) 0 From the continuity of A(x), B(x), C(x), and D(x) we are assured that unique solutions of these initial value problems exist on z(x) [-a, a]. Furthermore, + sy 2(x) + Yl(X) Y3(X) the function + ty 4(x), s, t being scalars, satisfies the initial value problem z(4)=A(x)z"’+B(x)z"+C(x)z+D(x), z(-a) The function z(x) will be a solution of sY2(a) If A Y2(a)Y"4(a) Y2(a)Y4(a) + tY4(a) # O, AI, z’(-a) s, z"(-a) (3.15), (3.16) provided s, A2 Yl(a) BI, z"’(-a) t satisfy Y3(a), a unique solution of the preceding linear t. 159 BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL EQUATIONS system can be found, and the corresponding function z(x) then is the unique solution of However, if A (3.15), (3.16). O, then Y2(a) Y4(a) y(a) y(a) p (constant). We can assume that p # O, because if p Y2(a) O, then 0 and by means of Taylor’s formula it can be shown that the solution of (4) Y2 A(x)-"’+ B(x) Y2 Y2’(-a) has the property + C(x)y 2 Y2 Y2(a) Y2"(-a) Y2 (-a) y2(-a) 0, contradicting the original assumption Similarly, p cannot be unbounded. Y2(a) 0 y(-a) i. Thus it follows that py(a), p < oo. (3.25) Now using (ii), and the Taylor formula, we obtain Y2(a) 2a + 2 a 4 [A(a) y ’(a) + B() y () + C()y2(a)] a < a < a, (3.26) 2a2[A(8)y y(a) (8) + C(8)y2(8)] + B(B)y2(8) a < 8 < a. On combining (3.25) and (3.26) we get i ap[A(8)y (8)+B(8)y2(8)+C(8)y2(8)] C[-a, a]. for all A(x), B(x), C(x) Y2"(8)’ Y2"’ (a) and y’(8), a 3 [A() y, ()+B()y 2 (a) +C (a) y2 (a) ]=i, In an attempt to determine y2(a), y2(), y2(a), we choose A(x) I, B(x) i, C(x) A(x) i, B(x) I, C(x) =-I, i, A(x) =-I, B(x) =-I, C(x) I, A(x) =-I, B(x) =-i, C(x) =-i, A(X) I, B(x) =-I, C(x) A(x) i, B(x) =-i, C(x) =-i, I, giving the system ap[Y2 (B) + Y2"(8) + ap[y 2 (8) + y(8) ap[-y’ () y2(),, ap[y 2 (8) y2(8)] Y2"(8) ap[-y 2 (8) y(8) y2(8)] + y2(8)] Y2(8)] + y2(8)] I i a a 3 3 [Y2 () + ...... + Y2 [Y2 () I a3f[_y2,,, -i I a a Y2"() 3[ 3 (e) y, () [Y2 () () + y2(e)] 1 Y2 (a)] I, Y () + Y2(a)] Y () Y2(a)] y () + Y2 ()] i, I i - J. WIENER AND A. R. AFTABIZADEH 160 ap[y i y () y2(8)] () 2 But this latter system in the unknowns a 3 [Y2’ (a) Y2 (a) y2(a), y2(8), Y2 (a) I We thus is inconsistent. conclude that A cannot vanish and the proof of the theorem is completed. Now, problem (3.14) with y(-a) it is easy to show that y(a) 0 has a unique solution. THEOREM 3.3. C[-a a], b(-a) Suppose a(x) g la(x)a(-x) _< n. Consider equation (3.14). b(a) 2 [a’(x) <- m a(x) O, and C (2)[-a, a] b(x) If 4m’n’2a 2 + 16ha 4 < l 4 then equation (3.14) with y(-a) y(a) (3.27) O, has a unique solution. PROOF. Since equation (3.14) can be reduced to equation (3.15), then by theorem 3.2, problem (3.14), (3.27) has a unique solution. Now, we consider the general equation (i.I) and prove the following theorems. THEOREM 3.4. Suppose f is bounded on [-a, a] xRxR. Then problem (I.i), (1.2) has a solution. PROOF. Problem (I.I), (1.2) =- + y(x) where G(x,t) (yl+Yo) (yl-Yo)X is given by Lemma 2.3. Define a mapping T:E + (3.28) C(x,t)f(t,y(t), y(-t))dt, Let M be the bound of Ifl on [-a, a] xRxR. E by I + (yl+Yo Ty(x) where E is equivalent to C[[-a, a], R] lYl i + a(Yl-Yo)X I_ a G(x,t)f(t,y(t), y(-t))dt, (3.29) a is a Banach space with the norm E max ly(x)I, Then from (3.29) and the estimates on lTy(x)[ i 1 a < x < a. G(x,t) and Gx (x,t), lyl+Y01 +a1 IYl-Y01 it follows that +7 azM’ 1 (3.30) and IT’y(x) I-< a IYl-Y01 + aM. Hence, T maps the closed, bounded, and convex B {y e E: ly(x) _< i (3.31) set lyl+Yol + a1 IYl-Yol + 1 2 a M 161 BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL EQUATIONS into itself. Moreover, since T’y(x) verifies (3.31), T is completely continuous on C[[-a, a], R] by Ascoli’s theorem. (I.i), (1.2), thus completing the proof of is a solution of T, which a fixed point of The Schauder’s fixed point theorem then yields the theorem. THEOREM 3.5. [-a, a] xRxR, Suppose f is continuous and for (x,y,z), (x,y,z) it satisfies a Lipschitz condition f(x,,)I If(x,y,z) where L I and L 2 > O. el IY-l < Iz-l, + e2 (3.32) (i.i), (1.2) has a unique solu- Then the boundary value problem tion, provided 2 a (L PROOF. (3.33) < 2. L I + 2) C[[-a, a], R] Let B be the Banach space of functions y [IY[ max[y(x)I, B B by (3.29), Define the operator T:B ITYI(X) TyZ(x) < a -a < x Then for with the norm <_ a. Yl(X) and Y2(X) g B, we have IC(x’t) [If(t’Yl(t)’ Yl(-t))-f(t’Y2 (t) ,y2(-t) Using the Lipschitz condition (3.32) we obtain a ITYl(X) TY2(X) IG(x,t) l[LiIYl(t)-Y2(t) + < L21Yl(-t)-Y2(-t) l]dt, or ITYI(X) TY2(X) < a2 (LI+L 2 2 maxlYl(X) Y2 (x) I, a < x < a, or [Iry l(x) Ty 2(x)[] B _< a2(LI+L2 2 [[yI(x) y2(x) Also, IT’Yl(X) T’Y2(X) ll B < a(Ll+L2) lly l(x) All of these considerations and inequality (3.33) show that T is a contraction mapping and thus has a unique fixed point which is the solution of (i.i), (1.2). Proof is complete. THEOREM 3.6. (i) Assume that there exist positive numbers m and r such that sup f(x,0,y) <_ 2_4ma2 2a (ii) f(x,y,z) has [-a, a]xRxR 2 r, for x E [-a, a], lyl <_ , a continuous partial derivative with respect to y on and J. WIENER AND A. R. AFTABIZADEH 162 2 f2(x,y,z) > m > -____n 4a [-a a], y for x g 2 g Izl R, < r Then equation (i.i) with the boundary conditions y(a) y(-a) (3.34) 0, has a solution. PROOF. a mapping Let B r C[-a, a]:ly _< r}. {y C[-a, a] T:C[-a, a] by (Ty)(x) B For y r and x [-a, a], define v(x), where v"(x) f(x,v(x),y(-x)), (3.35) v(-a) v(a) (3.36) and O. Equation (3.35) is equivalent to f(x,v(x),y(-x))- f(x,O,y(-x)) + f(x,O,y(-x)). v"(x) Then from Lemma 2.2, it follows that v"(x) 0 f2[x,Tv(x),y(-x)ld’r)v(x) (3.37) + f(x,O,y(-x)). Applying Lemma 3.3 and condition (ii) we have sup ]vCx) <- or from (i) 2a 2 n2_4ma 2 suplf(x,O,y(-x))I, sup lv(x) <__ r, x_< a < Hence, T maps the closed, bounded, and convex set B a <_ x < a, (3.38) a. r into itself. Moreover, from (3.35), (3.36) and Lemma 2.3 a v(x) G(x,t)f(t,v(t),y(-t))dt, a or a v’(x) -a Since If(x,v(x),y(-x)) <_ Gx(X,t)f(t,v(t),y(-t))dt. k, for x e [-a, Iv’(x) <-- a], Iv(x) _< r, ly(-x) <_ (3.39) r, then de. All of these considerations show that T is completely continuous by Ascoli’s theorem. Schauder’s fixed point theorem then yields a fixed point of T, which is a solution of (I.i), (3.34). EEOR 3.7. Assume that there exist positive numbers m and r such that (i) suplf(x,0,y) <_ (ii) f(x,y,z) has mr, for x e [-a, a], [Yl <_ a continuous partial r, derivative with respect to y on [-a, a]xRxR and BOUNDARY VALUE PROBLEMS FOR DIFFERENTIAL EQUATIONS _> f2(x,y,z) m > 0, for x e [-a, a], Izl y e R, 163 <_ r. Then the boundary value problem (i.i), (1.3) has a solution. PROOF. T:C[-a, Let B {y r C[-a, a] C[-a, a]: E by (Ty)(x) v" IYl <-- r}. For y E B r define a mapping v(x), where f(x,v(x),y(-x)) v’(-a) (3.40) 0, v’(a) + kv(a) hv(-a) (3.41) 0. Equation (3.40) is equivalent to v" 0 f2[x,Tv(x),y(-x)]dT)v(x) + f(x,0,y(-x)). Using Lemma 3.3 and conditions (i), (ii) we obtain sup lv(x) This shows that T maps B r into itself. v’(x) Since v’(-a) x [-a, a], hv(-a), then Yl --< r, zl <__ r, a <_ x < a. Also from (3,40) we have v’(-a) + I_ x f[t,v(t),y(-t)]dt. a Iv’(-a) _< hlv(-a) < hr. Moreover, If(x,y,z) <_ k for < r, then Iv’(x) _< hr + 2ak. Thus, T is completely continuous by Ascoli’s theorem. Schauder’s fixed point theorem then yields a fixed point of T, which is a solution of (1.1), (1.3). ACKNOWLEDGEMENT. Research partially supported by U.S. Army Grant #DAAG29-84-C-O034. REFERENCES PRZEWORSKA-ROLEWICZ, D. Equations with Transformed Argument, An Algebraic Approach, Panstwowe Wydawnictvo Naukowe, Warszawa, (1973). 2. WIENER, J. Differential equations with involutions, Differencial’nye Uravnenija 6, (1969), 1131-1137. 3. WIENER, J. Differential equations with periodic transformations of the argument. (1973), 481-489. Izv. Vyss. Ucebn. Zaved. Radiofizika, 4. WIENER, J. Periodic maps in the study of functional differential equations, Differentcial’nye Uravnenia(Iryazan), 3, (1974), 34-45. 5. HARDY, G.H., LITTLEWOOD, J.E., and POLYA, G. Press, London and New York, 1952. 6. BEBERNES, J.W. and GAINES, R. Dependence on boundary data and a generalized (1968), 359-368. boundary-value problem, J. Differential Equations, 7. CORDUNEANU, C. 8. USMANI, R.A. , Ine.qualities, Cambridge Univ. , Sopra I problemi ai limiti per alcuni sistemi di equazioni differenziali non lineari, Rendiconto Academia Napoli, (1958), 98-106. A uniqueness theorem for a boundary value problem, Proc. Am. Math. Society, vol. 77, no. 3, Dec. (1979), 329-335.