You are on page 1of 8
Methods of Math (b) ‘The eigenvalues of n periodic S-L system have multipl icity 9, 2.6.5 Self-adjoint operators and BVPs An operator A defined over a-lincar space of functions is called sey, - =< z, Ay > which in the present case is equivalent to int; “by Liv(e)[Ax(a)] de = f21Ay(2)] 2(0) dr where the functions y, z are supposed to be real, In case of complex funey a slight modification is necessary. ty S-L operator is self-adjoint For self-adjointness, =or-=0. orf? [yL(z) = 2 L(y) dx 0 From the Lagrange identity, we have Lotyt(2) ~ 2L(y)] de = pla) [(ye’ - y2)P But for periodic and regular S-L systems, R.H.S. = 0, or Se lyL(z) - 2L(y)] dz = O ie. the operator L is self-adjoint for regular ax periodic S-L systems, 2.7 — Tilustrative Examples and Exercises 2.7.1 Illustrative examples Example 1 For the S-L eigenvalue problem w+ dy¥=0, ¥'(0) =0, y'(@ =0 verify the following general results. poli. =! so = Next we Vaart Now VX must hav sin VAL Hence for 1B dn = [he eiger, 1un= Itis Fed, for me big) ba hase tn st. |e (a) There are an infinite number of eigenvalues, with a smallest but no las" | “cog (b) The nth eigenfunction has exactly n — 1 zeros, (©) The eigenfunctions are orthogonal and form a comple te set, Solution led ge, General solution of the ODE (D+ A)y = 0, D = dfdz with d > 0) can be nt hy, written as. X yacrcos VX x + eg sin VX i: Applying the given boundary conditions; Ply fg y(0) = 0 which gives 0 + e2VK1 = 0, wherefrom ey = 0, Soy =—-VAcq sin Vc Next we apply the second B.C., viz. 4/(¢) <0 and obtain -Vxq sinVXE=0 Yow VA # 0, because otherwise we woul aust have in VAL = 0 which gives Vil=nm, n= Id get a trivial solution: Therefore we O +1 £2. dence for the eigenvalues, we have dn = (7/0), n=0, 1, 2,» ‘he eigenfunctions are given by & for regular n= Cn 008 VXnt = cn cos(nmz/£), n=0, 1, 2, ') It is clear that ro = 0 is th eA, forn=1, 2, ) The eigenfunction corresponding to the n th eigenvaiue is given by ~1(2) = en cos{(n — 1)na/Q,n>1 e smallest eigenvalue. The other eigenvalues «++, Obviously there is no largest eigenvalue. (=) = cy cos(nmz/t), e010. +will now prove that Un(z) has exactly n —1 zeros when n takes the values 14 <=. n—-L **1"=0, yp = ay has no zero as expected. When n=1, tno argo "9 208(rr2/€) r Methods of Mathematical Physics “ 58 A nero of this eigenfunction occurs when 2/f = 1/2 ie. at x = 6/2, which lies in the interval (0, 2). ‘There is no other zero in this interval. Therefore the eigenfunction y has exactly one zero in the interval (0, €. (The next 2ero occurs at is at « = 3¢/2 which does not belong to (0, 4). Similarly, y2(x) = cze0s(2rz/é), O 0 and also y? > 0. 1) 2{0)] Hence it follows that A > 0. ‘This means that the given problem cannot have | negative eigenvalues. I also follows from (1) that \ = 0 if and only if y’ = 0 for all z, However from the given BCs it follows that the constant must be zero. i.e. y = 0 for all z. ‘Thus we conclude that A = 0 is not an eigenvalue of the problem. Therefore all eigenvalue of the problem are positive. This result has been obtained without = 0. Thetefyy actually solving the problem. We know that the eigenvalues of this problem are dn = (na/0}?, n=1, 2, --~ +¥0) Covamie3) 2(a) y'(a)} Express the function f(z) = z(1— 2) defined in the interval (0, 1] in terms of the eigenfunctions of the ST. problem: v'+¥y=0, v0) =0, v1) =0 vO) Solution v(0) yO) zt ‘The normalized eigenfunctions of the SL problem can be found to be tn=V2sinnez, n=1, 2, -- | Therefore we expand f(z) as 2(1 - 2) = CKAYn(z), where the coefficients can be calculated as follows. ) = ati), we 0 1 20 on va [ x(1—2)sinnmzdz lo a fon (5-3) oo | + eal vifos+0. (2 {samt ao }} nnd Be ns 0 ae A ethos of Mathematie! Piysleh - if ecomes which after simplification becom On substitution we finally obtain Example 5 1, 0<2S1/2 Express the function f(t) = { 0, /2 - Chapter 2. Sturm-Liouville Systems 63 ond R.HLS. of (3) 1 2 cet de = Gyn f 1 1 cos ayn) dle b 2 . tm ( stone) = om (,- tnt 2 20m ) |p _ tm (sun) On equating LHS and RHS above, we obtain 2am —Sin2am _ SiN yy — Am COSAm 4am a, ‘The right side can be simplified by using the second result of (1) which gives sin dq = ~(n/5) 608 on. In this way we obtain (—em/5) C05 Om — im 605m _ —6.608 am a, On From the above relations we obtain _ =24 cos om = =~ Dam — sin 2am Therefore on substitution for a,, we obtain =U coban fe)= x 2a, —sin2a, Non 2.7.2 Exercises 1. Under what conditions is the following B.V.P. self-adjoint? L(u) + Ar(z)u=0, 0<

You might also like