Chapter Variational Problem with Fixed Boundaries Functionals Euler's Equations Necessary and Sufficient Conditions for Extremum \soperimetric Problem Rayleigh-Ritz Method Lagrange’s Equation eoeeee 1.1 _ INTRODUCTION The calculus of variation is an important branch of mathematics and ~ -physics. It is a powerful technique for the solution of problems in dynamics of tigid bodies, optimization of orbits and vibration problems. The applications of the calculus of variation are concerned chiefly with the determination of maxima and minima of function of a single or more variables. Its object is to find stationary values or extremum of the given function, The problem of calculus of variation was first solved by Jacob Bernoulli 1698, but a general method of solving such problems was given by Euler. 1.2 FUNCTIONALS Consider the problem of finding a curve through two points (x, y,) and (2¥2) whose length is minimum. In general, we wish to find a curve C joining the points (x,,y,) and (%2,¥2) and having equation y=y(x), where y(xj)=y, and y(x,)=ya such that for a given function f(x, y,y" : J 1@-y-y))dx is a stationary value or an extremum. An integral such as above, which assume a definite value for functions of the type y= y(x) is called a functional.| 2 Calesslus of Varintions 1.3. EULER'S EQUATIONS The necessary condition for | F(x, yey") dx to be an extremum is that ! oF alo) ay # ay" This is called Euler's equation. y Proof : Let y=y(x) be a curve joining the points A(x,,y,) and B(x2,¥2) which makes 1 an stationary or extremum. yay + sno) Let y=y(x)ten(xs) be a neighbouring curve. joining _ these points, where e is small and n(x) is . (ooco va GDiEEar)y conn OWS cATeTENABIEL SS = Fig.1 function of x satisfying 1(>¢,)=0 at A_ and n(x,)=0 at B. Thén, the value along y=y(x) + n(x) Tey LSE) FeMa-¥ ¥en) dx This being a function of e, is maximum or minimum for €= 0, when > dite) Sen0-at 2-0 wvsse(2) = [otax—aray peta - eee [seae* ay de” ay! de | Since ¢ is independent of x,Variational Problems with Fixed Boundaries 3 =f: [Ero | ns atte _ de m, f° % yas (Sr), -f A(x} eas] aa . er “h [#-2(G) poe (4) [--70q)=0 at A and 7(xz2)=0 at B] Now, from equation (2) anid (4), we get a) which is desired Euler's equation. i.21.4. ANOTHER-FORM OF EULER EQUATION ‘The Euler equation is — ar 4(a)_ ‘which is another form of Euler equation. - 1.5 PARTICULAR CASES OF EULER'S j EQUATION The Euler equation is yale e4 HS 8 Calculus of Variat Every solution of Euler's equation which satisfies the boundary: Fconditions. is called_a stationary function or an_externallof. the problem. “The.stationary Function can easily be obtained in the following cases. mat of (Uf £ is‘independent of y, then , and the Euler equation (1) reduces to oy which gives the solution directly. - __ ~__ (8) bis independent ofy then mae and the equation (1) reduces to J=0 = 2 constant fufiction, whose ~~—-derivation-from-zero-is—very-small When -this ‘variation -is added to ... the. original. function. (x), we get. | new function y(x)+8y(x)i Then, fictional I ~gaints “an ~ increment, denoted by AI atid is defined as : AL SI {y + 8y)“Iy) 0) Lop gy Hoi20)-—10)—Variational Problems with Fixed Boundaries where O (dy)? = > OF S10) Roatan @) nf! dy” Therefore, { A(l)= 81+R; (y, Sy) veel) Now, let 5, n=1,2,3,..: be defined as variations in the functional, of first order, or second order, — oe S10, 11, 2,3, same Further, equation (2) may also be rewritten as AL =81+ 814K (y, By) ~ weeeee(3) where one more term in the Taylor series expansion has been considered and rest of the infinite terms are derioted by rerhainder -R,. Therefore, an éirient in furictional 1 can bé written as sum variations of increasing orders _. inthe functional. Lf-R,-is-very small enough to be neglected, then we have Aenea (6) ' which states that first variation of the functional is equal to the infinitesimal increment in the functional when the function y(x) (on which the functional * -depends) receives an infinitesimal variation. = _The above statement can be tnderstood as follows : respectively and have expression as (4) [serra frre meen ___. First let,_ » Iy)= i F(x, y) dx see) aaa b b Then, Al= f” Fx y+y)ax— [7 FE y)ax i aF , yy? Or nt bu ‘F FO t it Fey +8y)= Fo y+ bys 2 ay? Al= ae Fas [0 @y) Jax Jing (4), we can write : ses |” Fovy)dx = pe aE 5 re _——Farther, consider a: second functional of het form ef F(x, y,y')dx , where y' dyCaleutis'éf Variations ° ‘ b .s qhe, l= J) Fey +8yy'F8y ex — [P Fouyyyax xpansion for two variables, we have aE PO yt dy yt BEF Y, yyy 5889" Bi) )? + O(6y')" _ Using Taylor's series €: ren, MISSES Re v| OF oF here, 1= | | 5- ve wel(tay ¥ Me ay a Aflorgs +o SF +209) 67) EE | conse(13) order terms in 8y and 6y' and so on. nin sisptgai Necessary condition for extremum on a “variation must be zero forany-y— ~ ie, Si(y, Sy) 20 orsimply s1=0 oo - To show it, let the functional has an extremum (either positive or e) for some y(x). Let at first 8I(y, 8y)>0 for an increment, let k 8y, ‘placing k ec for-dy- in (1)-@),.we have or AT=OI(yr a) + “ (yk by) or Al=kL8L(, apt Ri (vy kéy)] ow, AI has to be positive for all small values (positive or ne; gative) of k =sgynall-so-Khas to:besimall). But the sign of RHS will be that ae Let k be negative, then second term in the bracket on RHS is negative, which is a i | _——sontrantictionEhus SH43',8y) cur. Similarly, we can show that &1 aaav nor be less than zero. Hence, the only possibility for functi i ey risgiven by 81=0. ty for functional to be 1.7 SUFFICIENT CONDITIONS FOR EXTREMUMS .. .. . “Here, we have two types of sufficient conditions, the first = t ones are deri in terms of the integrand function of the functional. Second type of fnalien agures Wat the tWoends-ofsthe-domain-of.the-functional_are cither:fixed or if one end is fixed then second fixed boundary ¢ aloes Rotor = occur ‘poundaries ii inside the given domain. . eS SS{ Variational Problems with Fixed Boundaries Higher Order Variations : We know that the increment in I may be defined as M=T(y +y)-1(y) wae Ley Soy of or Als B14 Gy) ay FO) —ally)+ When second and higher order terms on right hand side are neglected, the increment of the functional is equal to its variation of first order. Here, we use the following notations : ey) = 5 ~ ~1y), forn= 23, ees (2) We may write AL=BU(y) +8 Wy) 4.8 Uy) +a (2) where 871, 5, , 81 are respectively second, third, ...... and n variations of the functional I. Each of them is variation of the preceding variation, i _ 8 1=8(81) and soon. Also, Al=81+ 8 a (y,8y) veil (o) where R2(y,dy)= zi + = Pag). ween) | i = B lly) +. . Consider a functional Ig= [" Fey, yx we) subject to boundary conditions , you)=Yir YOa)=y2 ‘ ; Then, = jr Fouy+8y,y'+ By!) de— |" Foxy, yee Using Taylor's series expansion for two wattle we have 2 Fey 489,74 By IMO yy +B aay 7 : 1] (5,92 2F > sh = | +A) oy? Se 2y@ yp SE te a 5 | NOISE +9 z a 1 Then, 81 and 8% in (3) can be defined usually as in the article ° ied j——.. necessary condition. Gir t Also, the expression for second order variation is rewritten as 4Calculus of Variations 1p 2 aF 28 0 2 °F = J [oy a ONO oy = + (6y') aye © Sufficient Condition for Extremum Let us suppose a functional I(y) assumes a stationary value for a curve defined by a function y=¥ (x), ic, SU¥)=0, ic, the necessary condition for extremum of a functional I(y) be fulfilled. Then, for a particular function 7(x), we have AI=S14+R, (F, Sy) = 81 [If Ry is neglected] For the maxima and minima, we have the following conclusions : @ If P1>0 for every variation By (not equal to zero), then the function ¥(x) mintimtizes the functional Ky) . “GH IF-8°1<0 for every variation By , then F(x) maximizel(y). ~~ = There-is-a— of-curves-emanating-from—a~point—A(4, yi) and terminating at. B(x,,-y2)~- Amongst these-curves, there is a curve ¥(x) for — which the functional is either maximum or minimum according to whether 81 is negative or positive. Now, we will derive these sufficient conditions in terms of one of the —“three second order derivatives of F(x, y,y') which occur in (6). As at either — ----- point-A-or-B,-dy-= 0,-as-there is no variationin-y at anyones of thesetwo points, instéad, their values’are fixed in (5). = — But the slope on different curves at A changes, i.e, variations in y' (= Sy’) Not zero at Al or B. Hence, the requized sufficient condition for extremum will-be efine differentiable function-F,(x)-such that [Pe BER00 BYP] = [ CFG) (y)P+ 28, (8) By By" Fo) (yy [2 = 8) : (y);, = Gy), = 0, ie. no variation in y at A or B exist] ‘Adding (6) and @), we get a=" LF rth 1 +o 2 Sono irae EEVariational Problems with Fixed Boundaries 9 eR (ar ( #F J yp OP SE | =| ae, seeeee(10) J aay" i Then, (9) can be written as a f° ye BE Laydy +B 5 vy ¥ OF /dy'dy' This sation implies that the condition 71> 0 will iLimply that x0 and 81<0 will imply that 22—<0, ae 7 PY ¥ ¥ Hence, the sufficient condition for extremum of the functional (5) can be stated as Ps “If for some curve F(x), —“+— takes on positive oalues for all x in the interval w yr EXE xp, then the facto assumes mininnum value along that curve, Similarly, “Hffora ciurie F(x), that curve.” assumes all x, then the functional (5) is maximum along ay Remark. © The above sufficient conditions are called Legendre conditions. Solved Examples EXAMPLE 1. Show that the shortest curve joining tuo fixed points is a straight line. Solution. Let there be any curve joining the two fixed points A(x1,Y1) “and B(x),¥2)- (K.U-2005; M.D.U.-2005) Also, let § be the arc length of the curve. oy A P 2 Then, s=[" aS [dS = fax? +dy? | &,y, =f 14(%4) Js Po. : Ted —or S= t (1+y") ax 410 “Calcilus of Variations or 95 f £(y') dx, where fy)= (tty? ‘) The condition that the curve between A and B be the shortest path is that the integral equation (1) be minimum. ‘Therefore; the Euler's equation a 1(%) 0 sen(A) af = the constant b and a can be determined in terms of ‘which is the straight line, ints A and B as the curve passes through these a __—Hence, she shortest patho joining two points is'a straight line, “EXAMPLE 2, Show that the aren of the surface of revdlution of a curve y=y(2) anf" yay") + Kanpur 2010, Meerut 2010) ‘show that for AI eninima fhe curve. must be-cateny | Variational Problents with Fixed Boundaries n is. Consider an elementary arc ds at the| evolution is given by the curve rotate about x-axi any curve, i the curve. Then the surface area of r point P(x, y) of = an” y fary*) ax This has to be minimum. Since f=y Ja+y”) is independent of x, therefore the Euler's equation reduces to f-y' 2 = constant y or) -r SN (it+y”) | = constant | — yr) —y'{Zary2y?.2y'betsay) dy _ vO" =) sem yage 7 y ~ va+y") On separating the variable and integrating, we have JAR = [Bre —= pe c cosh” (2) = ~*8 or y=ccosh ( * 2) i c c c This is the equation of the curve which is catenary. EXAMPLE. 3. Find the extremals of the following functionals : j ) 4 . ii) a a oe io 2 y" @ Peryyd wf wed Gi FP arey rye :, 2 i (iv) £ Gf +y?—2ysinx) dx (@) 5 day dx ¢kanpur200) | Solution. (i) Here f=(x+y')y' | af which is independent of y, i.e., 1 az y | SoS2 Calculus of Variations .. Buler equation &- 4 dy ax a (af —|— |=0 2(3) o c or x+2y'sc or [a=$. oxy] 1 or Bac,-F or (ii) Here, wehave f=y"/x° which is independent of y, ie., st (y")_2y' dy" ay! eye “Ruler equation reduces to ~~ “Also; * (iii) Here, wehave f=(1+x"y)y' which is independent of y, Le. 5 =0. i pulley )y ia 14 y' The Euler equation reduces to 0,Variational Problenis with Fixed Boundaries (iv) Here, we have f=y? +y?—2y sinx a : Sf a2y-2sinx oy, a Also, S-=2y! ay Therefore, the Euler equation =-3(2) =0 becomes a 2y -2sinx-—(Qy')=0 y ~2sinx— F-(2y') or 2y-2sinx—2y"=0 or y"-y=-sinx I's AE. D? -1=0 which gives D=+1 CRaqet tae a Now, PL= -sinx) = Zsinx _Thus, = y=CR4+PL = qe%+c,e% +dsinx 1t+y? (0) Here, we have f= which is independent of x, ie, 2 =0. Of _ -2(1+y*) ayy? Also, Hence, the Euler equation Fl x d , 2+y?) |_ 1ty? _ dyY ity? ~ dy ity? a_(oy) _ ay _ - -(2) ©. ax Ce On separating the variables, we have dy dx +e or y=sinh(qx+c,)Calculus of Variations : “a4 EXAMPLE 4. On which curve the vintal= ff 2x" +Aycosx)de, y(O)= 0, functional y(n)=0 be extremized? (Kanpur-2013; K.U.-2004, 05) ) Solution. Here, wehave f=y®-y? +4ycosx * at a \ othat 2 =-2y+4c0sx. Also. =2y' ony 3 {I y «. The Euler's equation x(a). 0 reduces to da x - Ly = _. By Feo F2y)=0 or ~2y+4cosx—2y"=0 sy"y =2e08x-— ae .2cosx=xsinx “Pat Fegsinxxsinx nr (2) rations "y(0) = 0 and y(n)=0,-we get 9 =o and 0=c, cosm+c)sinn+ nsinn a “Therefore, substituting the value of ¢, in @), we get ye(qen)sinn chis only curve on-wi a ‘the'given functional can be extremized. EXAMPLE ‘5. Find the extremal of the function f (Gy? +12xy]dx with yO)=0-and~-y(1) =. (Kanpur-2009, 14; K.U.-2006) ~~ Selaitions We have-f=y2-12xy—_ at (1) at Sarx and S=2y! oy yoVariational Problems with Fixed Boundaries 1 or y"=6x or oy'=3x* +c or yar texten Using the boundary conditions y(0)=0 and y(1)=1 in (2), we get 0=c, | and 1=ltq +c | which gives | qa =0 | zz, Hence, from (2), y = x*. This is required extremal. 2 EXAMPLE 6 Find the curves on which the functional yee wi y(1)=0 and y(2)=3 can be extremized, (Kanpur-2011, 1 Therefore, the Euler's * equation yale us Js 0 becomes y x\ ay! 4 (-28 -$ (#2) =0 2° | Go Te where cis arbitrary constant, 1/3 3 —s v=(2) xo (2), le dx “te (2 ad x y= (2) Ata ~_ where cis arbitzary constant. Using the boundary conditions y(1)=0 aid y(2) =3 in (2), we get avs 4 yes 0=)- = 2 (3) vgtd and 3-(2) Sea.16 Calculus of Variations After solving ¢ 4 and d=-1. Substituting these values in (2), we get 2 y=%-1, which is only curve on which the given functional can be extremized. EXAMPLE 7. Find the stationary value or extremal of the functional nf2y 1 i (yy? +2xy) dx with y(0)=0 and y(e/2)=0 Solution. Here f= y"—y? +2xy ase (1) at Sa-2y+2x and ay ‘Therefore, the Euler's equation — — ~2y +2x-2y"=0 or y"+y=x or (D?+])y=x It's AE. D? +1=0 which gives which is required extremal or stationary value. EXAMPLE 8. Find the stationary value (extremal) of the functional 2 Ly anh a? —with=y()=0;= 2)s1_i._(K.U-2007; M:D.U-2004, 05) SoVariational Problems with Fixed Boundaries 7 Lity? a Solution. Here, f=} en x which is independent of y, ie, # 0 y ana HO {Viry" | ___iy' oy' dy'| x xyl+y? Therefore, the Euler's equation 5 - a( f)_y 20 or dx xjity? or yt=c2x? (ty?) or y*(l-cpx’)=ef x? ax oe ee YG-e2 x2) +e, eoeeee(2) 1 Using the boundary condition y(1)=0 and y(2)=1 in (2), we get or ont ft tes @) cq nde tee bh frac eg eee (4) oT From (3), 1 Aanegyse? or Gate =O tee ©) cy AT Now, from (4), : J Sa (-ey =z (l-4ef) t qech—2q=-d or nc -1=4-2, os © cy cy :. From (6) and (6), we get 4-2 =0 = =Calculus of Variations Substitute the value of c, in (5), we get c, Hence, substituting the value of c, and cz in (2), we get 2 or (y-2)2=58/1-% | or x? #(y-2)? =5 (y 3 vy which is required stationary value. 2 EXAMPLE 9. Show that the general solution of the Euler's equation for the integral (Kanpur-2010, 12) eG lax is yt ayt ee. Hence, the Euler's equation | ¢—y'2£|_ 2 <6 seduces to ax ax of . f-y5e c ~ where cis an arbitrary constant. 1 ayv2_y1__Ly" dey? -y? or AG+y2)?-y:— FY ce oy LY AY y vary"? ayy slo or y= pee vty? ay _fi-cy? a gy On separating the ‘ae yehive IntegratingVariational Problems witit Fixed Boundaries 15 fiuey? +h=x, where his an arbitrary constant. c -cy? or ATE Aceh? or Sa(x-hy ty? e 1 or (x-h} +y?=k? [ tske ak | ; EXAMPLE 10 (The Brachistochron Problem). A particle moves on a smooth curve, joining the two fixed points A and B, under gravity, starting from rest from A, find the form of the path in order that the time from: Ato Bis minimum. (Meerut-2006, 07, 08; Kanpur-2010) Solution. Take the fixed point A as origin and horizontal and vertical of the other extremity B. At time t, let v be the velocity ina “position P whose co-ordinates are (xy) and actual distance s measured. fromthe origin A. By energy equation, we have The change in kinetic energy = ¥ work done,in moving the particle Fig. 5 A to P. Therefore, ta(B)omer ) patra x Beyry) ds ds ==,2 dt= ae v8y v2ey Thus the time taken by the particle to move from A to B is m4 a i tf Bee yah pp er) Here, Pe 7 Ir __—which is independent of x, i.e, z. =0.o- =0 reduces to Wi (| \ 20 Calculus of Variations i = constant or == ~ ——afty(rehy 2 aoe sete eee or y(ity?)=2b, where bis constant. y=2beosty since y'=tany y=b(1+cos2y) d= coryay=cory(—tbeos y sin y or dx =- 2b (1+ cos2y) dy __ Integrating x=-2b (v + sa2e) 4a, where a is constant. oe OF ene Kab (By FE sin 2y)_ te “Thus the path is given by -parametric.equations J _________— x=a-bQ2y+sin2y) y=b(1+cos2y) AMPLE i Show ti hat thé Shotiestline between any two points on a cylinder isa Helix. -Solution- -betheline between pan ylinder, then _ a 1/2 | a anneeaees (40)? | 4 ni REE af do) a= snes i i bi a z- Q) where, 0, z are cylindrical co-ordinate, r = constant. Now, from eqn. (1), we get : oe arf ———_—Variational Problems with Fixed Boundaries 21 > f=[ler?@?y/2 to ae 0 88" 4x2 07)/2 From Euler’s equation, we have a _a(at)_, 30° dz\oe") © : a| Pe 3 o-—£| =" __|=0 ‘I. Bes] 5 a re : : iz 292 | gsc _ ol fraser | gl: Now, integrating, we get ——— By, eer re Squaring both the sides, we get wpe cee, tto@a (ltr 07) On integrating, we get “OS Cyz+C9 5 s+ (2) | which is the equation of helix. : z 2 Tfat p;,@=0,2=0 then co =0 | Letat pp; @=62 and z=zy then from eqn. (2), we get li \22 Calculus of Variations 18) any From eqn. (2), we get (182) : a a)? " EXAMPLE 12, Test for extremum the functional 0/2 | (a \? 2 Iyeal= f Te =¥? fax, y(0)= Solution. Here, we have Tye ie |@ 2)'-x] ax Bf al aey_ |—+-~-——-Here; auxiliary equations : m?+1=0 me=ti*~ Now, the required golution isVariational Problems with Fixed Boundaries w Bing th and (S)=1 > Acos>+Bsin5 | > Be=1. Hence, extremum can be obtained on the curve y=sinx. EXAMPLE 13. Find the curve passing through the points (x,y) am (x2, ¥2) which when rotated about the x-axis gives a minimum surface aren, Solution. Considler, Surface area = |? 2nyd8. 1 x 5 aS = Bay | ax s- |? Qmy1+y? dx = Q) ; f=2nyJl+y* «+ (2) (Shas to be minimum) 3 —a-~Sis minimum ‘fit satisfies Euler's equation. Euler’s equation is given by df, at) i - | f-y’)=0 axl? ay’) ( \ a{2mvi= y? -y’ (2ny) Ike” = ( 2 d az yy’ > apy tl ° ww On integrating, we get v2 yeiry? mes “where cis arbitrary cons! | ey? yilty?-y7I_ flty? mete oTCalculus of Variations J. > Squaring both the sides, we get = x+ec" cosh X46 coc wn +a cosh” tye ; xta 2__ysccosh—— € bs — id Cant be détévmined in Tans OF th ie CoordlINaTES OF TRE end point A and B as the curve Passes through these points. ‘yecanerte 14. Find the curve of least length joining tuo points in a plane, (Meerut 2008, 09) | Solation::Considér~ Hy ool= [ as ® ba yy)Variational Problems with Fixed Boundaries Squaring both the sides, we get » yredi(ty”) a-d)y*=4 dy _ Prk On integrating, we get —- yeextd which is the required straight line. EXAMPLES. Extremize 1 Ly (X= Jf Gxo¥ ~ye%) dx y@=L ylere Solution. We have f=xeY -ye* From Euler's equation, we have : at dat) ay axlay)°Calculus of Variations a Ye 0)=0 => xe KO = xeY eX =9 = xeY =ex x e Yo vas = x Now, taking log of both the sides, we get vlegentog{&) y=loge* ~logx _¥@=e-loge=e Te does not satisty given boundary condition, hence itis not extremal, MYM=L, y(e)=e-1 then it will be extremal, EXAMPLE 16¢Extremize Ly) cp Solution, Consider fay? ty’? From Euler's equation, we have af_ df at) ay axlay)=° => yex.” which is the Tequired extremal; EXAMPLE 17. A light trave] in a medium From one point to another point, so , ds : ° that.the-time-of -travel-given-by- _——— WERE aSAy eer 'etime-of travel given by~f. Vx, yy ete SIF are leageK ance ¥%, y) is the velocity of light in the medium, is maximum, Sincé.that path Of travel isVariational Problems with Fixed Boundaries Spit) Ba) ee ds Solution. Consider I= J ve %2, ary? a “ha Veey) Since fe oe “From, Euler's equation, we-have at af ae) =0 ay axl ay’ at _ df at) ay axl ay’) ae Hay’? av 4 y’ « ve “oy axl vfity? ¥ wr _qlt+y? av_|™ dy’? y” -y4{ v. fisy?] yay v. Tay? = 4253, 7 » “ary eR awisy? vader hay’ "a = 72. yy! ana 7pne ay ty? y- y firy Qe P: a Bo = 12: 291 =-(l+y7) a “aty)- vouey StewsCalculus of Variations % Za ty2 tM = vy" -yaty2) oO Oy ay IY YOY Na, a Now, vev(x, y) > Yes Putting this value in equation (1). av av = (te y2)evy’-y'4y 2 (l-y Yay vy" -y'(lty’ fee vy"=y" aryyZey? 5 abe es Lula pe tela a EXAMPLE 16. ii thé abe provionry orf ie natTre Of care, ————— Solution. If V=y .Variational Problems witht Fixed Boundaries 29 Now, equation (1) becomes py seep? +1=0 > P dp+® 1+p* y ° On integrating, we get fe Flog (1+ 22) +logy =A, where 2 is constant = log(1'+ p*)+2logy = 24 > Jog (1+ p*)y? = 24 —— Onintegrating, we get xtb=f Eo ” Let ay? az dy=—taz > ydy=-5 ait @__pge_f2_,2 x+b= 3/ eee a! y sy (x +b)? =a? -y? > (x+by +y? <2? ‘Thus, the required path will be an arc of acircle.Calculus of Variations “EXAMPLE 19° Test for an extremum of functional ‘1 20 T=y (l= [) Gy+y?-2y°y') dx yY@)=L yAya2 Solution. Consider f=xy+y? -2y’y’. From Euler's equation, we have . at_ afar) S- +; )=0 ay ax lay’) : x4 2y ayy 2 (dy?) oO x+2y—4yy’+4yy’=0 xt2y=0 Now, ~ Hence, there is no extremal, because it does not satisfy given boundary conditions. trong TEE 20. Find the extremals ofthe Functional and extremum talue of the ) Uy@le ho 2 e2aua nf © TYOM= [ijn xy Fax witrcomttion-y| CO MCN [F (e=y'Pax subject ylO)=0, y2)=4 ‘Solution. > “Tyce fis yay /: = @ Here fos2yr2Variational Problems with Fixed Boundaries 2 On integrating, we get xy’ = constant dq . 29y ie, ax c dx dy => > iy 2 On integrating, we get y= ~£.44, where dis constant of the integration ... (2) x { Now, (3) =1 = -2e+d=1 yQ)=2 => -ctd=2 On solving, we get c=l d=3 w @) _ From equation (2) and (3), we get — y= -443, which is the required extremal y "Now, extremum value We have 1 +3 x = 4 yaa 7 x = From equation (1), we get | aft y2y2 2 hie y ax ari ® tO wifta x = ya x 4 ix op 24 fap _ _ =f z -| thet? 1 Hence, extremum value is 1. «@ Tiy@l= fp @-y Pde 0)f=(x-yP From Euler's equation, we have ay dxlay’)= Calculus of Variations d. From equation (2) and (3), we get “Whiclris the required extremal:—-——-- ~ Now, extremum value aX _x oe yea | -—-——~ —— yee -—Now, fromrequation (1) 2 = Jp &-y'Pax ee Eo dx ——Variational Problems with Fixed Boundaries 2 = f, dx =[x} =2. Hence, extremum value is 2. EXAMPLE 21. Show that the variational problem of extremizing the functional 3 Z Hy dl= f; y@x-y)ax 1 7 y@)=45, y(=1 has no solution. (CSIR NET-2012) Solution. Wehave f= y(3x—y) => f=3xy-y? From Euler's equation, we have ee ar_dfat\ Su)“ \o0 oy ax ay’) Z d > 3x—2y-F(0)=0. = 3x-2y=0 , \ 1 EXAMPLE 22. Extremize Uy (9]= Jy. (siny + cosy)ax i n is, y(0)=0, yM=> i Solution. We have f=xsiny+cosy From Euler’s equation, we have — a _Afat\_o sey ixhay’) => xeosy-siny-2(0)= 0 i > xcosy—siny =0 \ > xcosy =siny LF tany=x y= tan x,Calculus of Variations y(0)=tan0=0 4) = tant Za y()=tan*F — whicitis thie required extremal. EXAMPLE 23, Find the extremal of the functional b LyQ= {> (y? +y’? +2ye%) dx. (Kanpur-2009) Solution. Wehave £=y? +y’? +2ye* ae 7 ay 729 te a —. Therefore, CE. =cye* +e NOW Pd tk.4s Variational Problems with Fixed Boundaries Hence, complete solution is y=CR+PL yacye™ tone * _ EXAMPLE 24. Find the curve, the time taken along which the least, when velocity a | Gaya} | “any point of itis v=x. 09) Solution. Consider ~~ Blyol= Now, f=f(x,y From Euler's equation, we have at_ af at) ¥ axlay’) bees y \ : d a altar : y’ \ => 0 woe fisy?, z) On integrating, we get t = constant = cy xfl+y? Let y’=tant, => Be tant ek _ = iit xylty? xsect 1 sint x +e", 2 vex i II36 Calculus of Variations > x=cysint E qe 4 ¢ Now, gy dy dx dt dx'dt stant cj cost dy a > ae t apc csin On integrating, we get y=—cy cost +¢y > Yc, =-c4 cost From equation (3) and (4), we get 24 (yep)? 2G sin? 2 YY —¢g)" sezsin® t+, cos’ t = 2 sin? ostt) Pa(y-ay act which is the required extremal. EXAMPLE 25. Extremize the functional swear (@-»] y)=t~ yGaj=i. Solution. We have - oo 2 -(Z x = fay?-y? From Buler’s equation, we have dann > ~2y-F2y')=0 ego @) (CSIR NET-2011) ©Variational Problems with Fixed Boundaries 7 > oy +y=0 -Here, auxiliary equation is m?41=0 > m=ti Solution is y = Cy COSX+ Cy sinx __Here y(Q=1 = y(2m)=1 = cy cos2n+cgsin2n=1 > 140.021 => cp is arbitrary gal Hence the required solution is -_ yscosx+cysinx. © EXAMPLE 26. Extremize the functional Ly (x= m (y?-y? + 4ysinx) dx. Solution. Consider. f= y’*—y” +4ysinx 2 oy 44sinx > From Buler’s equation, we have at _af a) _ ay” dx ay’ ; dann = By +Asinx-F-Qy')=0 . (4 ion 2: a ee = -y+2sinx-S{ 2 ay > eas 2sinx. oe yr > (D? +1)y=2sinxCalculus of Variati Here, auxiliary equation is m24120 => m=si CF. = ¢,cosx+cz sinx Pl. Now, 2sinx 1 => J sinx dx=-cosx Hence, complete solution is yaCRaPL Pe tee, yy")dx Y=C1Cosx+ co sinx —cosx Remark. @ This is well-known brachistocluon problem which derives its name fror the Greek words 'brachistor' meaning shortest and 'chrons' meaning time. 158—FUNCTIONALS-DEPEI _ DERIVATIVES INDENT-ON “HIGHER-ORDER A necessary condition for fo be extremum is ---(1) afd fag) d? fae ) dy ax lay!) ae layr JF? ~~ -Let-y = y(x)-be-the-functi boundary conditions, fon which-makes (1) stationary and satisfied the Y¥Oa)= Yas YO—)=¥o, y'oq)=y; ‘and y'ea)ey; Also, let n(x) be differentiable function such that Nba) =0=n(x, ‘Now; replacingy by y+ 2) and WO%)=0=7'(xQ) (D, we get en s-s(2)Variational Problems with Fixed Boundaries 39 1(@)= s £(x,y ten,y'ten!y"+ en") dx rn: 8) To find the stationary value of (1), we find the stationary value of (2) for dle) _ de 7 e=0. But Ie) will have a stationary value when We write f=f(x,y,yy") and F=f(x,y+teny'reny"+en") :.,@) becomes I(e)= J” Fax Differentiating under the integral sign, we get aie) _ ps ae ype (af, aE, aE de da de % = Sy (Sy ayt P aya OHO * RO <0; when e=0, which gives i iS at, ae st 1N'+T,; dx =0 ; I & ay! Hn) * i—— Integrating by parts, we get 2 af af |" _ pe ar at af a - ax+| 2 ge 4 faves -t aShucl eo gp i ma 2 (af pg ae "or giz 2()- = (= ) oo 0 [By using ()] f Since this equation must hold for all values of n(x), we get ¥-2(8)-a(8)- In general, a necessary condition for the _functiona x I= j * fx yyy", yl) dx to be extremum will be ma a df(a), @ (a n af 4. 3()- (8) neon Ee) which is known as Euler-Poisson equation and its solutions are cal i—extremals. Mle40 Calculus of Variations ——___—. Solved Examples EXAMPLE 1. Find the extremals of the followirig functionals : @ Tly@)= E (16y? ~y%+3?) dx (i) Ifo(x)]= fg (2xy +y)dx (Kanpur 2014) Solution. (i) Here f= 16y?-y"+x2 Al ‘The Euler-Poisson equation is x aS) Se) 0 esseeal2) : at aE = 232; , #9 and [=-2y" ay ay? ay" / “Therefore (2) becomes a Py gWtaa 32y—2y" =0 (D* -16)y WEAR. D*=1620" = (D2=2)(D2+27)=0 won ee DER, £91 Hence Y=Ce™ + cpe™ +.cy cos2x + cy sin2x Hes which is réquii i jon is “ad (at) Ga “ay dl ay!) ae lay") aise — eS and ay ‘Therefore (2)-becomes & 2x a (2y"), =0Variational Problems with Fixed Boundaries a d’y stex dx® ae a 5 Its solution is y = ata Fc_xt + 05x9 + yx? +.C5x + Cy which is required extremals. 4 EXAMPLE 2. Find the extremals of the functionals k " (tay? +x?) dx that satisfies the conditions y(0)=0, y'(0)=1, y(n/4)=y'(n/4)=1/V2. Solution. Here f=y"*-y? +x? seen (LD) The Euler-Poisson equation is 5 eal Se or (D'-1)y=0 = (D?-1)(D?+1)=0 3 D=+1+i Hence y=, e* +c,e +c, cosx +c, sinx y'= ye" —cpe™ — cg sinx + cy cosx Using the boundary condition in (3) and (4), we get yO=qtate = GtQty=0 oa, ©) YO=q-eaty = Gretyst (6) (2) =e" se" ca cet tee t/4 +e ae 5 sees) cet/4 —er/* pote “5 ; soeeee(8)Calculus of Variations After solving these equations, we get S=c)=e=0 and cy =1. Hence, the required extremal from (3), we get y=sinx EXAMPLE 3. Find the extremal of the functional Iy@l= |" Ly)? 29? +y? -2ysina]de. PNET(CSIR) Dec 2002] Solution. Here f= y"*-2y"4+ y?—2ysinx eared a The Buler-Poisson equation is of dfot), &@ fa sels Sl-0 nn, (2) 5 ala) ala)? a or. 2y—2sinx + 4y"42y* =0 or (D* +2D? +1) y =sinx . It's AE.is____D#4.2p?. 7 1 os, “+04 *Fixed Boundaries 2sinx . 1 == «| on [: oa 1 variational Problems with 22) XZ sinxt 2| 2 = -Xsinx- 4 (Dp? +1) asinx __X in ory 4 x : (Dp? +1) 8 - Hence, y=CF+Pl <1 eo yal +c,x) cosx + (C3 +04X) sinx = sin* SeapeyeaT AE Afhich is required extremal. EXAMPLE 4. Show thal the functional it the curve which extrentizes me, ce (y®@—y? 422) de tnider the conditions yooyet, yO=0. WR/2)=0 and y'(m/2)= (Kanpus-2013; K.U.-2007; M.D.U--2005) 1. Solution. Here f=y"-y¥" + x wee) ‘The Buler-Poisson equation is ot (a) f(a <0 altar at a at Bf _ ny, a0 and 5 =2 By OY By ayy ‘Therefore (1) becomes a -2y +-S(2y")= y eeey) 0 It’s AE. is D'-1=0 => D=£1,4i Hence, y=cye* + @™ +€, cosx+cy sinx dy tos Xe ~ i y= 3K eye" ~ ce — cg sin x Feycosx eee (3). “Applying the given boundary conditions in (2) and (3), we get 1=y)=4 +e +6 i. = OF y'M=c.-c, Hcy O= y(u/2)= ce"? + ce? +ey= Calculus of Variations “Ls y'@/2)= qe"? — ce7/2 05 Solving the above equations, we get %=G=c=0 and o=1. Hence, from (2) y= cosx EXAMPLE 5. Determine the extremal of the functional 1 =f) (Suve+ry) ac subject io y(-1)=0, y'(-1)=0;y(l) =0;y'@)=0 Solution. Here f=Luy? ‘hy The Euler-Poisson equation is a alae) alae) ‘Therefore (1) becomes Ce a ay =0 oy * A+ ae (ny") or + a Tts solution is ,_dy Ae are al HQQX+Cp ae (3) ‘Applying the boundary conditions inv (2) and (3), we get 4 d=y-n=-4F oP 6 P —cgltey — $3.2 —2egltey-- > gral Hegl? teh +4Variational Problems with Fixed Boundaries 45 ae a = 50, =e and cee Alu pa! Substituting these value in (2), we get a gd 22 a Yar gy 2x? +1) or y=- se Ge PY? which is the required extremal. =~. EXAMPLE 6. Find ihe extremal of the functional £ (y+y4dx that satisfies “fie boundary conditioris. yla)=yor yb)=ya, y’(a)=yo. y’'(b)=yi- (K.U-2007) Solution, Here, we have ‘b J, Wty") ax fayty”. em oe ot ay’ ay’ ay Now, Euler's Poisson equation is given by 2 Ha (at), #2) <0 1. Q) dy dxlay) dx? dy’ Putting the values of onau (1) in equation (2), we get ~- ros “= 0 = 1=0 + (which is not possible) Hence, extremal cannot be determined. EXAMPLE? find the extremals of the functional Ly ool [? ty"? 720° +97 ax (Kanpur 2009) Solution. Consider faly’"P-207'P +97. Now, the necessary condition for existence of extremal is a afar), da) ay ax ldy’) * ga lay”)Calculus of Variations 46 = ay- Se ayy Sa Tey" o dy # 7 v9 +2—| ae tr = yaad aelae!-° nay dty 4+2—5+—7 = 0 = we at aty dy —4+2—S+y=0-- - > = ad ae Here, auxiliary equation is = m* +2m?+1=0 = __. @m? +1? =0 = m? +1=0, Hence, the required solution is Y= (cy + ax) cosmleg + cgx)sinx which is the required extremal. VARIATIONAL" (CASE: (SEVERAL ‘ARIABLES) _A necessary condition for pe ee @ 0°68 ar extremum is that #4 Heo, 12O120n 9... 4Q) Ut SaYaroyYn sally the boundary conditions ¥ 64)= i Yi(%9)= Yu —- -We- ‘ider- Nr (X)/ No (X)orNe(X)— are arbitrary--functions stich that &) ‘on the boundary. YuYarerYn bY Ya tern T404)=0 Now ‘i 11 Y2 FEM Yq + EaTeVariational Problems with Fixed Boundaries a7 To find the stationary value of (1), we find the stationary value of I(e) for € 82 9. _.A(e) will have a stationary value when ae) _9 Meo WO .o eo Fey, 0, ene We write afl Yu Yar Var Yarn) and. F=£(x,yy +21 Y2 + €2%- Yq EM Y2 FM +) Therefore, from (3) and (4), we have Ie)= J" Fdx [Eee at 6) ")- Differentiating (5) under the integral sign, we get dle). pu OF “(= Pa. ae 5 dx = =—”At d 35, I ae pa + Spm Jo ale) _, 3e; 3 =0, when EF Bs ..=0 which gives. integrating (6) by parts, we get, See oe Jon ax=0 [Since n, (4) =0=1: 62)] ate) Similarly =0, when & =€2 =. which gives = fd {Of ‘ aye 4 by, =0 and s0 on. : hich is the general case of Euler's equation,48 Calculus of Variations Solved Examples dt dt x(0)=0, y(0)=0; x(1)=1.5, y(1)=1 is stationary for x=1+t? /2, y=t. ay? EXAMPLE 1, Show that functional [ [2(4) (¥) le such that (Kanpur-2011, 2013) 2 i : Solution. Here £=2x+ (3) # (2) i dt dt The Euler's equations are a ae)” ax dt Lax" | | |e : a5 Ff oF - 2 Bene a RG nO Therefore (1) and (2) become F pasty Px gH O8 Fae dad - E ‘i id and -£ > The solution of differential equation () and (4), we get e Fhcitsh-cy,-where c and cp are arbitrary constants y=bt+b,, where b, and by are arbitrary constants Applying the boundary conditions in Sand (@)- we get- 0=x(0)=c, = “= Dati ti = egxjcosx+(esteaxsinx ay ©) ay , ec, +eqx)sinx + (cy + ¢4x)cOsx + €, cosx +c, sinx Also, ey =—(C, + eax) cosx — (cs +¢4x)sinx—c, (2sinx) +¢4(2cosx) Therefore, from (3),.we haveCalculus of Variations, %y 2c, cosx Bort AY = (Er +C9x) cOsx + (cy + cyx)sinx—2ey sin x + 2cy Hence, the required extremals are Y = (Cy +0gx)cosx + (cy +,x)sinx \ and = z= (cy + cyx)cosx+ (cy +¢4x)sinx — 2c, sinx +2c, cosx | EXAMPLE 3. Find the extremals of the functional | n/2 dx)? (dy) . § {2»+(4) +() dt with conditions x(0)=0, x(x /2)=—1, y(0)=0, y(n/2)=1. 2 2 Solution. Here f=2xy + (2) + ch vest duasnyreaianne dt i ar at . i Ry hae | 3 af dat 3 <-Sy" |= d : | ox dt eee om i d i Ss =0 i 2y- Gx) and~2x To solve the simultaneous equations (1) and (2), we get We differentiate (1) twice, | I's AE. is ion (D4 —2)=0-->- - (BD? = 1)(D* +1)=0 = De+1Lsi. | Hence, x =cye" + cne*+ cy cost + cy sint From (1) a ® GPx tape pgitb inp ate pmeree y= a2 =cje rope =o; COSTE F, cy Sintariational Problems with Fixed Boundaries 51 Using the given conditions, we get O=x(0)=c; +c + C5 He O=y()=e, ey Cy -L=x(t/2)=c,e%? +e, L=y(n/2)=qe%? te, - cy Solving these above equations, we get sae: cy =) =¢3=0 and cy=-1 These values stibstituting in (3) and (4), we get the desired extremals xs-sint, y=sint 4 EXAMPLE 4. Extremize t Hy), 2001= [0 (y’? +2/ +-2y2) dx =n/2 +g y(0)=0, ¥(3) =1, 2(0)=0, (2) =-1 2) * 2 Solution, Consider fay 42/2 4 2y2 + We know that the necessary condition for existence of extremal is af df ae\_ ay axlay?)“° ; ard at mt Bale)“ “From equation (1), we have Seayy= 4(2 22 PVI=0 = Tax ar iy n9 =» 482), j, Wz and By ~ Gg 22')=0 > A (#)-y, ie, ae? ; low, differentiating equation (2) w.r. to x, twice, we have dty dz at ae i dty ‘4 oy so axt (B*-y=052 Calculus of Variation: Here, auxiliary equation is m4 -1=0 (m?-1)(m? +1)=0 = m=#1/ti Thus, the solution is y=cye* +cge™ +03 cosx +c sinx + A! ' @ 7 . | Therefore, Saget * _o5 sinx-+c4.cosx 1 | 2 : | a’y = = i ' = eye 4096 * 6 cox~ Gy stx -- oe d Now, from equation @, we get 2, we : =1€% +.c9e7* —c3 cosx— casinx From equation (4) arid (6), We get a — = cye™ + cpe* —c3 cosx—cy sinx z= cre? ego eg 1, cy $¢2—c5 =0 | wt On sdlving aqiaor (6; tyreyana pMEeG —— ¢=0, cy=0, cy=O and cy=2 Putting these value in equation (4) and (5), we get ‘INDEPENDENT VARIABLES Theorem : The 1ecessary conditions for existence of extremals for functional “ds Sb F(x, y, 0, ty, wy) dx dy oo Q aF_ afar) afar) da axlaa,)” ay duy)™ =0Variational Problems with Fixed Boundaries 53 where u(x, y) is continuous and has continuous derivatives upto the second order and is préscribed on the region of integration D. (Nagpur-2009) “Proof. Let u(x,y) be extremal of (1) and ii(x, y) be neighbourhood of Su(x7y) such that U(x, y)= u(x, y)+en(x, y) (2 : | where € is small parameter ahd 1 (x, y) is an arbitrary function such that n(x y)=0 -- 8) onthe boundary C of D. heZin af E (Projection) Now from equation (1), we get . Ipa(x, y= ff, Fey, OW, iy, By) dx dy I()= Wo F(x, y, U+en, Ux + eT, Wy +eNy) dx dy = a oF oF oF 1©=ffp [Fr yeu, aeuptene ren aE aon, Ze n00e| dx dy = x Yy (By Taylor's Theorem) ‘The necessary condition for existence of extremal is 21) a i_ C 0 7 ( \ | ie, Slo [sna dx dy=0 \ flo 75; Fax dy fy (nz 8) dx dy =0 -@Calculus of Variation Now consider, (ar) (ar) af ar) afar) a "Me Was “alae a" "alau,) ~ "ay(3u,) (a. ar) i, (™ 3a, au, | ey a -2q- d =f, [Zin-n ay nex y Now, equation (4) So ngeeey~ Soy sof _ of aR) af ar) ID 9) “39 Bus, dx dy=0 Thus, n(x, y) is an arbitrazy function. Therm, (BF_ afar) “af ap) <0 | du axl u, Poel ay) we have x = R w- 2m) eh, )=0 which is tequired necessary condition for existence of extremals,Variational Problems with Fixed Boundaries 55 Hence, this equation is known as Euler-Ostrogradsky equation. Remark. ‘@_If functional (u(x, y)] contains derivative of higher order two, ie. T[u(x, y= SI F(X, YoU, Ue Wy Unger Uxys Uyy) dx dy - _ Now, Euler's Ostrogradsky equation is : a a a? ae e -—-ks - 5x oe Yo gy Fay) ae Mae Seay Buy + oye =0 2 EXAMPLE 1. Diricilet’s Problem : Find the Euler-Ostrogradsky equation for Ifuce, y= Jf (2) (2) ax dj “Y= Mp hax! *ay. jodie value of are prescribed othe boundary C ofthe domain D. i ‘Solution. Here, we have == au)? (an)? _ 2.2 Furey me y)=(2) (2) Sug tuys . ) From Buler-Ostrogradsky equation, we have ax_2(#)_2(}.0 Sa ox ldu,) ~ By (Bay ' stichonnste as | from equation (1), we get a asain af (2uy)=0 0-2: @u,)- : a 2 (a,)=0 p=? Fle Fy y) + 2 33) 3 faa) 9 7 2 *3y ley. j 2, 3 au eu = ax? ay” => v7u=0 ves required extiemal (X, Y)- 4 which is Laplace equation and it gi56 Calculus of Variation EXAMPLE 2. Fittd the extremizing function for Tue y)= ff, (sy (2s {Za} Bey) ste] ma where £ (x, y) is known function. Solution. Here, we have £LG6 YW Uy Wy, Ugg thy oa Y faa? (92 at u ww) = + +2 ~2uf(x, y) (3) 2) Laz) (ax =) G fauky tupy +2u2y ~2uf (x, y) oD) From. met onen equation aR J+ 6, +h, )=0 fy eu, ast Os) ay Go * HF Fayy => —£(x, yrs * (au) eB ( ow a8) a? (au) a Bxa5 axay) “ay aye) omsrarea EXAMPLE: 3. Find the extremizing fonction for au)? , (au\? | Hho dito tay ry) |dedy ) is. known function. “Here, we haveVariational Problems with Fixed Boundaries Suz +uy +2uf(x, y) From Euler's-Ostrogradsky equation ar 2(at) af ar) au ax\du,/ ay\au,)=° Now, from eqn. (1), we get a 2F 9) -Fey)—Fe ay) = =0 au) a (au) ax) aylay)=° fo y-2(* au au => a tal (x, ax ay? (xy) gives the required extremal u(x, y). Ee mn Se ier 2 “e) +2uf (y, alas dy dz £(x, y, 2) is known function. £28 (8, Ys 2) U, Ug, Uy, Uz) (2) (ys (28) saute y-29 +uy ¥u2+2uf (x, y, 2) 3--From Euler-Ostrogradsky equation at 2(2}- afar) 22) du ax\du,) dy da,) “az Bu,)~° afau)_ afau) afau £00 ys 2)— a(®)- aylay) 7 2(@)- 7 au au atu Sat Sy age LOYD which gives required extremal u(x, y, z). =u2 57 .@) -@58 Calculus of Variations © 1.11 ISOPERIMETRIC PROBLEMS Tt is necessary to make a given integral I= [" tGy,y)dx subject to the : constraint J= |" (x, y, y’)dx = constant bs Such problem involve one or more constraint conditions. This type of problems are called isoperimetric problem. . . The necessary condition for integral {” Hdx to be an extremum is sy ) = 0, where H=f+Ag, 2 is called Lagrange multipliers. We determine the Lagrange multiplier 2, together with the constant of integration arising in the solution of Euler's equation and integral J having given constart value, 1. Solved Examples ——————____ EXAMPLE 1, Find the extremals of the | Tytel= J? y* de given mat [i ve ny Solution. Here £ oa a talannon eda isoperimetric problem (Ku. 2006) 7 + We write ; a H=fag=y "say Now, H must satisfy the Euler's equation ~o2. BB ae dy dx o oH ay Therefore, (1) becomes 4-—Qy! Integrating Ax-2y'sa oy dx-a dx? _ ax pena seoramnnae fiche eea 59 Variational Problems with Fixed Boundaries EXAMPLE 2. Find the plane curve of fixed perimeter and maximum area, (KANPUR 2010,13) Solution. Let ! be the fixed ; périnieter of a plane curve between the points P and Q, with abscissas x, and Ex, (see figure). Also the area between the curve Yeaxds is ; fv sense) have to maximize (2) subject to constraint (1) Hig. 6 _f=y and g=J/(i+y?) i i : SeSHeterg-ytalary Now, H must satisfy the Euler's equation oH -3(2- 0 ‘Oy dx ay! 1-4} 39" _ | ax! Ja+y*) | x-—4'_ | lary") | 2 “ay or (x-a)? (1+y®)-22y? =0 y ; (x-ay 19y___(x~a) } T= Gena) a [i a]Cafculus of Variations j Integrating, Y=VPP—G=aF]+b or (yb) =2? (x -aF or (x-a)? +(y-b)? =22, which is circle, EXAMPLE 3, Find the extremal of the functional T= fF oP yate una the conditions y(0)=0, y(x)=1 and subject to the constraint ff yar=2. (Kanpur-2014) 3 Solution. Here f=y?-y? and g=y, | «We write” _ > = H=f+Agey?-y? +ay Now, H satisfy the Euler's equation i dH a é oH -2(= Ht) -0 a ae i i Therefore, (1) Becomes 2 i tyta-Leyy=0 or -2yeh-2y'=0 2 | SOv th gy" y tha 2y | or ~oytey=h or (D?42)y=2/2 i i “Wes ABs -D? + 4 CR. * cy cosx + ¢ sinx Now, Bia} | ows Di412 =a+pty7h = (1-D?+..), nI> " nia 3B 3 spina vis dea Hence, yng cosK Fey sini 2 | _+ sneha tounry condition fh en gat O=y()RQta/2 = cH-a/2 =o gah a) Solvig @) and wees daetVariational Problems with Fixed Boundaries 61 —sposx+e, sinx + i 2 7 2 Now, from constraint i y dx =1, using (5) aC 1 JF [ eos + exsinx 4 Jex= 1 . i [ -Jsinx—c, cose =1 2 21, x lin - 2c tori or easy Substitute the value of c, in (5), we get i y= a-cosx) +2 2-a)sinx §:which is desired extremal. [SS =EKAMPLE 4. Fhad a fiction y(x) for which [° (x? —y)dx is stationary, — jo F given that {fF ax=2, w(0)=0, y(1)=0. - : 4 i a: | jf - Solution. Here fax*—y# and gay’. . | We write . } _ H=ftaged -y*+Ay? { 4 i 4 ——The Euler's Precio is au _a(aH), - ay ax FF au { and 2H oy \ a Therefore, (1) becomes j 7 a -S-2y)=0 or Dy+2y"=0 Dry - G2) y +2y) y’+Ay=0 or (D7 +2)y=0 It's AE. is D+a=0 => Dati ¥ =C, COSAX + Cp SiN AX Using the boundary conditions y(0), y(1)= 0, we get pene ccissaneys anne emtstik rans] > WSR NCL AAD : gsCalculus of Variations OByO)=e 3) | 0= yl) =¢, cos +c sind soee(4) ! From (3) and (4), we get sink=0 or sint=0 [Note that c) #0. If cy =0 leads to y=0 , we have sink=0]. sind =sinnm, where n is an integer. or A=nt Substituting the values of 4 and c, in (2), we get ¥=c sinnx Now from constraint f° y*dx=2 and (6), we have 0 i cf sin? nx dx =2 or Laos or co =42 Substituting the value of c, in (5), we get - qe + — ¥Q)=+2sinnex,-wherenis an integer. EXAMPLE 5. Prove that ‘the tremal of the isoperimetric Problen: Tye fl y2ae with yayes, (4) =24 subject to the condition [* yax=36 is a parabola: a Sélittion. Here f=y? and gay. We write H=f+Agqy*+ay The Euler's equation is enc acrs> suena kaZariational Problems with Fixed Boundaries 63 r-deyyeo or yt=t _ Integrating, z yiagxte Integrating again, ye) -t¢ FHC, eee (2) Applying theo soriition ¥O)=3, (4) =24, we get 35 y=keq +e 24 =y(4)=40 4 4c, + cy Solving these above equation for c, and c,, we get and cp =A-4 a Sabai te value of c, and c in @,we get ——y®)= oky #4(7- Bera 4) we) “The value of f(x) in using the constraint { y ax=36 tes fi [$e +(r-B)na- #) |ax- 36 Fes fp _ BR) x i [Be “(7 SD pro-as | = 36 joe) ayeqanpniat ca Te 1 7, 9 ; ghey ok had i ibstituting the value of 4 in (3), we get i y(x)=x? +2x $whiclris parabola. ; i EXAMPLE 6. Find a curve C having a given length, which enclose a maximw 24 3 ** Solution. Consider, area bounded by the curve C 3 A= . (xdy— x) | “ale @8y=¥29 — = le (x ZI) ay [ x dx | {Calculus of Variations 1 : : =ahe (xy’-y) dx ‘ Mi Now, the length of C 3 sa, \l+y? ax wa (2S Since, I=A+is =f. Gy’-y) taf. fry? dx =f (Bow yeasty?) dx Let HF 6y'-y)+% he aH_ 4 Therefore, 21-2 oy. 2 1 oeceomemrpet cence tatomemnns dyer eve tA: horas 8h ele IMR on HNe Hime SBATREY sRifiaticial Problems with Fixed Boundaries 6 On Integrating, we get where c is arbitrary constant. Squaring both the sides, we get 922 ee ay? = (x0)? +(x y)2y? Soy? DP -(-0)7 = 6-4? 7 (x=«1)" (K-47 (x=c1) ? — (x4? =) dx fa? -(x-e4)? y?= Now, integrating , 7 gir 1) pF Ka) yeas + 2 1/2 __ m2)? =? —x(x—ey)? 2 242 2=> (x=) +(y—c2)° = éptesent a circle. “Hiénce, the area will be maximum. SSEXAMPLE 7, Find the extremal of the functional [> y/?éx under the stiaint-{> yax=1 given y(0)=0 and y(2)=1. ee Hef+igey” +ay + @) . _Now, H must satisfy the Euler’s equation — o> ala’)66 Calculus of Variati ae a From equation (1), we get a-Zey) =0 = 2d(g)-0 2d2y _ x oH and >;>=2y ay > fy.* On integrating twice, we get : ho J VE text - : ge taxte where cy and cz are arbitrary constants. Now, y(0)=0. = c)=0 y(2)=1 => A424 =1 1- B i rong flere ee aes nasea isu suf eaters epee els semua uence mi = poem amma fas EER retesettional Problems with Fixed Boundaries é : 1 & yQ* hich is the required extremal. = EXAMPLE 8 Find ihe solid of maximum volunie formed by the revolution of a Given surface. 5 Solution, Let OA rotated about x-axis and passes through the origin. g yy 5 P@y) B of Hi : @o) IK (@.0) z § Now, the surface area enclose by the curve y=y(%) is & a i S=[, 2ny ds g z S= fp 2myyity” ax () = nd volume V= [* ny? dx -Q) ? i 0 fany?.g=2ny1+y [From Lagrange multiplier’s method] H=f+ig “ 7 He ay? +22ny flty? « ice,-the necessary condition for existence of extermalis H-y’Hy = constant ‘rom equation (3), we get 2nyy’ ny? +h.2nyylt+y -y’x = constant _ hay? ny? +2nyyl+y”? — tee1 Calculus of Varia > ny? + any vit y? Since, curve passes through the origin (0, 0), then we get c=0. Now, from equation (4), we get 2nyd ny? + =0 > Now, squaring both the side tey?2& yet On integrating, we get § ydy > an? —y? =x+cy ___It.is.given that.curve.passes through (0, 0) , a -2heey acai i Piast anne dtbAbid lela es bbb ts reins ielnoneee one ceial Problems with Fixed Boundaries 69 belle: cy =-20 is __From equation (6), we get [92 =y? = (x—- 22)? (x-2nj24y? = 49? Hence, the required extremal tepresenis a circle. LE 9. Finid the extremal of the functional - pty 1), tay A=f, gh y-y sat eyed aan egral constraint - te [aod ap fe ax ty? =L lution. Here, we have ty-y'), gayx? ty - (D) “Hed oy-y04 hs i? 4 y'? ae) the necessary condition for existence of extremal is oH 2 (28) = = ax dt\ax’)~ . aH d(aF\_ Se alay) 7° -++ 3) uation (2) and (3), we get a{ 2x! maT \ .. 4) + 5)Saver 6 0 Calculus of Variatic f On integrating equation (4) and (6), we get : yt eae => y-c, ah @ yi ar “€ ra ty? . [02 +y? i and shy i vuemtinanssy ee fama see cnet ay’ x SSF =n > x-Q= b x’? 4 "2 . x! 472 Now, squaring equations (6) and (7), we get (52 4. y/2) 2. 2 2) X + X= Cg)? + (y—cy)* = A2] A FY (xe) + (y~ er) (Fay?) (<-e2)? + (y~ey? = 22, which is the required extremal. 1.12_LAGRANGE's EQUATION : p EB g 5 3 z z g 3 a ; z 4 3 3 a i g ohne as at_time t The Lagrangian functi -where-T-and-v-arehinet ‘and-potential-energies respectively of systemis a function of q),q3,....4, and 41 Gar q, at time t. By Hamiltonian’s principal [Lat is stationary. mews Therefore, Euler's equation must hold good, aod ( Z| = i te, +L = I 99; These are called Lagrange's equation which determine the motion of the | system, erg | EXAMPLE 1. A _— of a light spring having constant C, is | sef into vertical motion. Find the Lo ~-equation~of ‘istion of tie inass by using Lagrange's equation. INET (CSIR), June 2003] Solution. Let the displacement of the mass m from the equilibrium position A toBisx, attimet, “Theti, thé kinetic energy of B is gesun, sreeepenyfx:oewmcepernin rm mene 2s ete71 Problems with Fixed Boundaries 1_ (dx T= am (sey Now! the work done during its fell me we ff (mg C0 dx= amg-$00 nin foe » Vis the potential energy of the mass at B, then ve-w=3ce —mgx fore the Lagrangian function cy=4m() amgx-Lod = hint amg“ ZO" Seer the Vagiange's equation is yl fi fey.y dx formed by replacing the inde; pendent variable or by a sinmultaneo™ ment of the required function and the independent variable. Ther. 7 a femals of the functional from the. Euler's ¢ uation for the formed integrand, * Let the independent variable x,y be ch (u,v) and y=y(u,v) ¥ sgetiowy St72 Calculus of Variato and |** * #0 Yo W JiGuy.ypaxe f {> vy yerwy BZ] xy vy) du Xy tXyVy | eRe @ | which satisfy the Euler's equation | i aR d (a : | arava)? ~ & Therefore, the éxtremal of the original functional are determined after the | Wanefermation, and by using the transformed Euler's equation (3). Solved Examples" 3 a itdneumatepurmnent wn ig Tyas [PY (ety? ey?) de. Solution. Here f=e™ yey? a romania nee ses ewes baleen A ayte*)=0 eee) or ; bo thd (2) ‘or sie x=logu and. y=u reduces the functional to B . Twa) ef wtavt)du vo |triational Problems with Fixed Boundaries 73 Revt_y? oF oF that S=-2v, Satay! that So=-2v, So=420 "The transformed Euler equation 2-4 ( 2 J 0 becomes av du -2v-Leavy=0 du mm ov'+ve0 =0 D?+41=0 gives D=ti. we v=c,cosu+¢sinu “Hence, the solution of original functional is y =, cos(e”) + ¢, sin(e*) [slogu=x so that u=e*] EXAMPLE 2. Find the extremal of the functional 82 fe? +42) do, where r= 16). (KU_-2004, 06; M.D.U-2005) 81 2Soiution. Consider I ]r(@)]= fp? ye? +1?) do =e 1: Let x=rcos®, y=rsin@ =xt+y? => rayxtey? : ‘aid— tano= 2. Since S --rsinds Foose ; oa) : a ar vena Srareosar Gain wn Q) “Now, squaring and adding equation (1) and (2), we get GQ) eg eto (SFB =f Vly? &7” Calculus of Variations Let us assume that at @=@; => x=xq andat @=0, = x=xg pi VF +r? do~ f fi+y? ax cM ek amatsne ar ARR Lis (O)l= tly l= f? ty" dx Now, from Euler's equation, we have » af y\eg al fy?) => 2 tke ity? [i= Squaring both the sides, we get [ooetaneapo racer rane bleenpcned moms: senvconyes nixon d mel: rape a On integrating, we get Y= cyxtey _ Hence, .extremals rsin@=crcos0 +c. “1-14. VARIATIONAL PROBLEMS IN PARAMETRIC FORM __Let the parametric representation of curve x=O(0, y= vio, tyststy Then, we have t I [x(t y= Sie (Lx, hy) dt ti -where. desh denotes the derivatives w.x.to tand integration is along (1). 38 I eC ibosgsgae : | evaiiationial Problems with Fixed Boundaries ar 5 If the, value of functional (2) only the line (1) and not on the number of | ways in which parametrization, it is riecessary and sufficient that integrand in * (9) does not contain it explicity and I is a homogeneous function of first degree pain’ ay | Therefore, we can write Pos toy y kx’ ky akiOx yxy’), k>0 i = Consider a new parametric representation as =x(0), y=y(t) and t=n(t), [n'(t) #0] LS. of (8) becomes E490. 0-¥ Olat = [YOO OO rOMOE yy — [= x=x(0), se Xe oe, 2x =x, (D.1(0. Similarly y'=y_ (0.70) © Since, > isa homogeneous function of first degree in x’, y’ 50 Lx (), y (1), Xq (4) (8, Ve OW (HY= (0) (% ¥, Xt Ye) ee : . “IE Gy. <0-y Wlat= [2 otto) yO. x val [- t=n(t, C= (Ol ol) presentation. { Hence, the functional (2) must be of the form Tbe. yO ff Sey VIE fe > is a homogeneous function of degree one in x” and y’. sre yea ; ay’ Putting these values in equation (1) and (2), we get f , ( » \ Si leg ang SA)’ jg “at 4 7 +y? . dt 1 3°24 yr . 7 On integrating,-we get... __ 7 : s X= =A and ; v We +y? Now, dividing Hence, which is straight line ina plane. EXAMPLE 2. Find the extremals of the function te v2 4 2 Ie yyx ty” dt. ~ Solution. Here,Euler’s equations are_Widfiaiteithl Problems with Fixed Boundaries sale) ax dt ax’ of _ df af) ay atlay)~ wae x! +y? of OF LP ay2 x ay x 30° 2g y2y Way ax’ 2 ty Way! . (1) 2) a) =. 4) w=Calculus of Variation. > On integrating, we get J Bz Jase, cyeoshY = x40, Ge 2 | Is J (-y? +x*y?/2)dx is minimum at every point of the domain, | subject to the conditions that its first order variations at the two ends of | __.... the domain are one and zero respectively, is given by : _.8 2 ° x 14° the general solution of the extremals for the functional r 7 5 log (e*y?xe*y*)dx under coordinate x=logu,y=v. ] we ane = “7 4 | (Show. that the extremal .through_origin_for-which—the-funetional transformations | @) Show that the extremal of 1= f Gy-y")dx, ¥(0)=0, y(4)=3 is (=x) /4, 1 | |iitional Preblem with Fixed Boundaries in a lf Determine the extremals of the functional T= [ (Sys ¥y) de _ subject to the conditions y(-1) = y'(-l)=y(1) = y'@)=0. «Find the general solution of curve for which the functional ) Y oxy ty”) dx is extremum. 2 : oe how that the extremal-for 1= f/ le) + (2) |e dy is solution 5 “of Laplace equation V7, z=0. i). Discuss the Legendre extremum conditions for the functional tm f (822 aay oy) 2(0)=0, where z= dy. ow. that the extremals y(x) in the semi-infinite domain (0, ¢) such Wat (0) = 0, dy(0)=0 and dy(c) is any finite value and for which a functional of integral (x? +y?+y") has constant minimum value at ‘very point of the domain is y=1—exp(-x) - - Find the extremals of the functional f° yy/9dx with conditions SUyC)=1. Extremize the functional Tye Jo" (y? +2yy’—16y?) dx 7 . = nr) _ ee Es y(0)=0, vf) =1. ixtremize the fanctional Vives fF by? +6xy +98] dx eS O=0, y@=2 | .02).;.,Bxtremize the functional ee Viy@l= Ss, (y’+y") dx.80 Calculus of Variation (13) Find the curve on which the functional Hy (l= [}. Gxy+y?-y?)dx y)=y(1)=0 can be extremized, (14) _ If the extremal of the functional ir wae with y(0)=c, y(2)=B is Parabola passing through origin, then find the value of « and B. @5) — Extremize the functional =f? 3 (2) , dx. ma Ca) (16) Find the extremals of.the functional : when y(£1)=0; — y’(e1)=0 @?) Find the extremals of the functional 5]? Jisy2 42? ax x | — | “8) Bnd nett (8) for which f (cory hicieron EERIE = 1 1 fp ydk=2, year, | => 8) Provelthat-the Biler’s equation for function j J faeyy? +266) Yy' bey" dx {8 a second order linear differential equation, (.D.U.-2008) ~~ @0) Find the extremalof the functional & (9? +29? 4y"2)ax (21) Find the extremal of the functional Po ayyrax. Boundary aha ina iational Problems with Fixed Boundaries 81 @ how that the geodesis on a sphere of radius a are its great circles. a (M.D.U.-2005) i23) Show that the extremal of the functional a 1/2 Bog fa) 4e(2) ight H ty at at dt is a straight line. (M.D.U,-2005) 5 ANSWERS * scopes . 2 °° 40. y= sind We 2y=P-2x 1D yeer tex 20 5 Bay E(x») 44.0=0,B=1 15. yee +09 msds 2 2 bs x? 1 ee 47. y=ax+b and z=cx+d, where a, b,c, dare constants x2 2 SinmAx 20. y=(Gqtene+(Gteyxle™ 2% y= “Tie ‘problem of calculus of variation was first solved by J. Bemoulli in But, a general method of solving such problems was given by Euler. Specessary condition for [* £(yyzy')4x to be an extremum is that 4(2) 20, which is called Euler's equation. " ily possibility for functional to be extremum js given by 51=0 oF ay ay" £x Sx, then the functional assumes minimum value along for some curve (x), takes on positive values for all x in the rval 1 ( 7 ) =0, where H= f+g, Ais alled Lagrange's multipliers. y “od x ~ Lagrange's equation determines the motion ofthe system.Problem of Geodesis8~ suppose We have Soxfare — ® Specified by a Vector equahan teat tuy) : he shastest cuive Cof minimum Rength) tying on the Busfece « and camected two pets A ard B 4 Bwiface = 13 cabled the geodesic Comnecting the wo bomtss The equations fur the geodesics of a ante the Ewen Equations of the Cassresponcing yeriationa) pstobl em amedy— the problem of finding the yninimum distanw bfuw two poivets of the surface os FULER'S EQUATIONS ON GEODESICS F- A cue Dying OM she —-swifa He Hus) — A) Can be Specified by che equations Usuttd 9 VP= # beng @ postamelus « lo cl a = ow = Fy Fe Hyde cated the Co-effrerents of, the w4)—O) Gre Hy HH — (3) Whee quandihes ante fisat fundamentad form of -aunBace a- Re anc Rength bw she Boils © AL») so the povrameten 4, B gwen by 7 tun = “RETR RaN a OH) Falex!s equations for the firctimad (4) ane 2, [Tew I~ del Se Sea BFA Fa0") <0 25[ Ewa euvtiour|- a. Fa fewszrae ravi \ . me and. Ble) cerowpardgFiamblel Find the Gioderica of the cixewan Cynder @ He (aos 8, a SIn@, 2) Si Now Hoa (acaso, asinorz) Then Sip = (-astho, acoso, 6 Oh= Coron) ‘|r @ These fore 5 — (1) E= 5 Sh=ar Fs Jig: Sh =0 [-e Ge Hy =| “The ase Sength blow the hus points ACL) and Blo) Lying on the ie lia CD is given by +he Runctimag Me.z]- 4 Eo 2Felz +Gz™ bt —— (+4) on Vez = a Oey Be dt —— (3) Ewlen's equations > functional (s) aie = wos legos | — (6) = d.- ae z! — (1 © pens -° > “These suits On Intepiatins Yields eel ag, 2! Jae 2> Ja*giag> & Dividing +he Secmd of -the epuation by the fists We obtain 2) _ ce = constant ea = dz =C ad Which Rag the AolvHons Z= Cetd —C8)Equation (2) siepreserb 0 +tuw0 posrarr eben family of e Pebicat Dines Lying on the cylinder C). Tuas & Qeodesics on Cylinder co is a heltr- ramble 2. End the geedescis of the Sbheie H= Casing w3xpy9 ASno sinby Aca3xe) SI" we have Fl =Ca sine carpy & Sine Sindy C030) then ai = Cacass Coady Acare Sb, -asin6) Gig= Casing Sings aSing carp 9 0) Thexefosie, — <4 7 as F=sig: Slo = CoS © Cas ot & cate sirrgtesm o =a Se eh oa 2 Dia at Fe Hy iy = —c® sino caso carp Sing +a gino care Sing > =O Gz dig Sg = OP Sin’o srg +a sire cash = ase —the anc Length Joining the ve points ACL) and Beds) Lying on the Sphere cp 13 gwen by The BancHonal steele JUF(gey 4 apg) (aby 42, : : = JRE tae eB bet “pEereaye ie q = * a Jagr save dh a sN “re o+snmo dd » where @'= “3 Here, Fe Nevis OJGF Suze Which 1s Independent ob ¢ Aor Cwvrespording Euler's Epuchion ts > Alea — -6 aol = constant IO sin = 2a, ‘ = OL Sin*o~ OL rwhere ¢ 13 Some Constave J O74 576 => So _¢ = sn'o = (04 swe) JO=sn*0 3 Cot= srie- sro 3 c’et= Siro csinto-c*) > Co! = Sno Jsiro-c = de = one (site-2) = nk T= cere? ‘ ap 3 do, = Sie Ji-c*meco 0 $ Cc adh _ Cemeco_ _, dh= Ccae’o 6 de {-c* case l-cceree”O = C Cameo © Tite guccting » we get ' = carl / Coote Jae d= Cox cot } = ($-c) = cos ( cee} Sa 2 Cws(p-c')= Clete ioSlas(h-c)= CecatO- wheae ¢.-_6 oO JT-c- = Cos cose'+ Sind snd= ¢ cos, SZ asmoosd cascl+ aq sina sidsinc = ¢ ace = otcascl ty Snel = Az > es 4 sme! =. =) Ax+Oy=z, Where A= caac! and B= Sine! The This 1B the esucchion 4 dhe plane passing fhewvg the Cente Corie) of dhe sphere and Ivcessecting the sphere afeng a qaeck Chik. Hence, geedeaie on a phere & Om aC of a sre ack asich. Example :- find dhe Jeodesicz an he gol). The Jeodesica an he plane of the ferchimat Cy] = “[ FRE an As Here, Elooy y) = Jt ty® which otoes met Corman y empliately, Se) Cosrcsponding Euler's epucch'm 3 BE es Costant ay The implies oY ic 2 ye cc uey) Plane, ig an exhemal Teg SY Cc J =yl- Ar whe A= J lege a Ye Au+BThis 18 @ €fuection of the Shaighl me in © the Plane: Thun Geoclesics In he plane ane shaight Ants: