You are on page 1of 218
10 Asymptotes 10°L Definition A straight line at a fixed distance from the origin, is said to be an asymptote to an infinite branch of a curve, if the perpendicular distance of a point P on the curve from this straight line approaches zero, as the point P moves to infinity along the curve. 10°'2 Determination of Asymptotes The ecauation of a straight line not parallel to y-axis is of the form, ypmmxte = (l) _ Excluding at present the case of asymptotes parallel to y-axis, it is obvious from-(1) that as x approaches infinity, m andc must both tend to finite limits for asymptotes to exist. Let p be the perpendicular distance of any point P(x, y) on an infinite branch of a given curve from the line (1), then _ © P Vien If line (1) is to be an asymptgte to a giver curve, than as x00, pO avy Lt (y—mx—c)=0 x+00 or Lt U-mx)=c (2) x00 Also from (1), we have Y= = emt Taking limits on both sides as x—;00, we have a)= _c) )= ut (+) EL (mtg im xte . m= ut (+) --@) . zee 248 ASYMPTOTES 249 Thus from (3) and (2), we get the valucs of m and c and hence the equation of the asymptote. Note. A given curve may have more than one infinite branches then it is possible that cach branch may have separate asymptotes. Hence a given curve may have more than one asymptote. Example 1. Find the asymptotes of the curve Ft P= 3ax*, ‘Sol, Here the cquation of the given curve is 2+? = 3ax® (1) In order to determine the asymptotes, we have to evaluate m and c given by Lt (y/x) and Lt (j—mx) respectively, then x xem y=mx-te will be the asymptote. Dividing (1) by x°, we have 3 i+( x) a) x x Taking limits as x co, we have 1+m*=0 ° Lt Of)=m se =(1+m)(14+m*—m)=0 “ m=—1, as the roots of 1-+-m?—m=0 are not real. Now c= Lt (y—mx) x+e = Lt (y+x) te m==1) xoe Let y+*=K, such that as x0, K>e Putting y=(K—x) in (1), we have 238+ (K—x)P=3ax* or 3(K - a)x' Dividing throughout by x*, we get 3K? , K? 3(K—a)—=+ 420 Taking limits as xo and Kc 3(c—a)=0 or c=a The asymptote to the curve is given by y=mxte Le. yooxta or “p+x=a is the required asymptote, Note. The method used to determine asymptotes in the above example is not convenient. The following methods are much easier and quicker to obtain the asymptotes. 250 ENGINEERING MATHEMATICS 103. The Asymptotes of the general Algebraic Curve Let the equation to the curve be (dox™ bayx""'y fagx"-2yt + ... any") 4- box"? + byx" 2p “EB ax"~8y? $b by") + (Cox egxBy Ht cox" AVP oe HF eny) + oe = 0 + ee (1) Further let »=2x-+c be an asymptote to the curve (1) Since each expression in the brackets is homogeneous, (1) may be written as vibes (Z)4etea(Z)tora(Z)4-=0 # where ¢, (2 ji is an expression of rth degree in ae Dividing (2) by x", we get (3 }S #e a(Z }S tea(Z)4n0 ++-(3) Now taking limits as x00 : and am (2 Jem, we have | Folerd 0. +-(4) which determines the slopes of the asymptotes. Substituting y=mx+c in (2), we get xn ( (m+) +2" $02 (m4 ) bx bee (mt £) 4-50 Expanding ¢n (m+ ‘ dn-a (m+) ete. by Taylor's _ theorem, we have at [bomtS da! (od + gh S 4" m+) FSH ony 4... tar [ baa ot £8 na +5 os Beaton)... ] 20-t[ beset SH" oat ya 8m atm) 4, +-=0 Arranging the terms in descending | powers of x, x*balmn)-+ x" Leda’ Gn) +$en-vGn)] +x? (S bn" OM) ECF tn ym Beaute) 7 . ASYMPTOTES 251 tant s balm) + Se TP nalen) +8 nam) +$n-a(m) ]re-co, we have $n) + }n—(m)=0 or cH a [provided #',(m)40] +6) Now cquation (4) is of the nth degree in m, giving n values of m Say 71, M72,.. , mn. The corresponding values of ¢ Say Cy, C2---» Cn are given by the equation (6). Hence the asymptotes are =m x+e, YHmox+Ca, etc. 104, Parallel Asymptotes In case $,()=0 has two equal roots say m. = then #' a(n) and $,-:(771,) both become zero. Now om Om) Thus c takes the indeterminate form 0/0. In this case c is obtained from equation (5) of the Art.-10°3 as 57 bn" (1) +c$'n s(n) nan) =O ['nGn) =¢n-x0m) =0] chee we obtain two values ofc say ci, Ce vorrespondirig: to. m=my (a repeated root). Hence the “asymptotes are y= mxter yHmxter, ie. paralle} asymptotes. Similarly in case there are three parallel asymptotes, c’ is obtained by the te equation - Sr eos a fon) rotten) 4 bal)=0: "This being cubic i in cwill.give. three” values of c, correspond- ng to three repeated values of m. 252 ENGINEERING MATHEMATICS 10°5. Working Rule (1) Substitute y=mx-+c in the equation of the curve. (2) Equate the coefficients of two highest powers of x to zero, (3) These give m and c and hence the asymptetes. Example 1. Find the asymptotes of the curve, 284227 p—xy?— 2p" xy — p> 1=0. Sol. Putting y=7x +c in the given cquation, we have XP 2x°(x+c) — x(x c)®#—20mx-+t c)? +x(x +c) —Gnx+e)?—1=0 +2n1— nt? — 2m) x? + (2e—2me -- 6rr"'e + m— m")x*+...=0 or Equating co-efficients of x* and x* to zero, we get 1+ 2m—ni— 2n = (14+ 2m)6m+ DGn— 1) (1) and 2e—2me—6m"c+m--m*==0 or = om oat a From (1), we have m= —1,1 From (2), we get (@ when (ii) whea (iii) when Hence asymptotes, are . yea y=-x-l and pox or 2ytx=1, yrxt1=o and y=x. 106. Shorter Method The polynomials ¢,(7), $n-s(m), ¢n-i(m) etc. can be easily obtained by putting x=1 and y=minthe nth degree, (n—1)th degree, (n—2)th degree, etc. terms respectively. -The following examples show that this method gives asymptotes much quicker. Example 1. Find the asymptotes of the curve Pax V— 2x +.2x°—7xy +3? +2x* + 2x+2y+1=0 [A.M.LE. 1974, 75] Sol. Putting x=1 and y=m in the third and second degree terms we get $3(m) =m? —m—2m?+2 and = $Xm)=—7m + 3m? +2 The slopes of asymptotes are given by . $4) =m—m—2m'+2—0 pvewss RULES oe : 253 or Oto 1)(m—2)=0 Now & m=—1, 1,2. Again 2 (mn) = 3m? — 4 —1 . © =~ tlm) or @ when (i) when (ii) when - or ThA 42x aay b 2p — x Oy F2 Sol. Putting + : Putting x =1 and y =m in the third degree terms, we eet a Bo OM) = Le Ana 4S 2md sO 2 On=2 +4m+42mt #1 (2) =—1-9m, The slopes of Asymptotes are given by a og Me) 144i 522m =0 or Qm+DGn+1)?=0 ” m=—1, n=—1, -1 Now when m= _ ts) ba! (mn) Hence the asymptote corresporiding to m= —} is » pecdxy-l or 2y+x+2=6 when m=—1, ¢ is obtained from the following relation 3 c $3" (m) + eFa' (mn) +4, (an) = 0. “dal (m0 LY Fema (10-412m) +e(4-4+4m) +(—9m—1)=0 2 (5-6) te0+(9=1)=0 [ gts. ” cH42V2 c= mel] 254 ENGINEERING MATHEMATICS Thus the asymptotes are yorxt2V/2 and = -y=—x=2V2 or yrue2/2 and y-+x+2V2=0 Hence asymptotes are 7 2y+x+2=0, Ytx=2V2 and y+xt+2v2=0. Example 3. Find the asymptotes of yi 2xy*$ 2xty —x1— 3x84 3x pL Zxy?— I — 2x2 . +2y°—-1=0, Sol. Putting x=i, y= in the highest degree terms, we gct . $4 On) =m — 2m +2m—1 ° bs (m)=—3 43m +3717 —3m? 2 (m)=—2+2m"* $, (m)=0 do (m)=— 1, The slopes.of the asymptotes are given by $4 (m1) = m1" ~- 2m°-+ 2m —1=0 or Gn? — 2) — 2m On? —1) = or Qn? —1)n?®-+1) — 2m? — 1) =0 or (mt? —1) (nr? -- 21 F1 or Qm—1)(m-+1) G2 — 1 or Qn-1)3 (m-+1)=0 or m=1,1,1, —1. . When m=1 (thrice repeated roct), ¢ is obtained from the equation : Ste ot Fp dat (omy te te! (om) +4; (m= 0. or = 4m —12-+ FS. (618m) | $y" (on) 4m 6m? +2. t+e(4m)+0=0 bs” (m)=82m0°—12m or co (4—2)4 c? (3—9)+ 4c=0 | $4’ Gn) =24m —12 Com=1) or 2c? 6c? ++4c=0 Also or 2c (c?—3¢+2} $a’ On) =3-+6m—9m? or 2¢ (e—1)(e—2)=0 $9" (in) =6— 18m and . $ (m)=4m Hence asymptotes arc y=x, y=x-+1, y=x+2 “ ¢=0,1,2 AsYMPTOTES 255 For. m=—] = — Salm) _ = FG) =0 (for m=—1) Hence asymptote is y=—x or y+-x=0 Thus asymptotes are : YEN, Y=R+L, yaa, ytx—0 EXERCISE 10 (a) Find the asymptotes of the following curves: 1. xty-bay*txyty'+3x=0. (D.U. 1979) y— Gxy* + 11x*y-69 + y+x=0. (D.U. 1977) 3. oxy —sty tP+x*—y—-1=0. (D.U.. 1976) 4. x94-2x"y—xy®—2y?+xy—y'+1=0. (D.U, 1975) 5. x8+22°ytxy?—x?—xy+2=0. (DU. 1974) 6. 3x°+2x°y—Txy*+ 2y"— l4xy+ Ty? 4a+5y=9. 7. 8. 9. YH yP=x+y. xy—xy+axyty*+y—1=0. Poxy +2xy+y+1=0. @ : 107 Asymptotes Parallel to the Co-ordinate Axes So far we have been excluding the ‘case of the asymptotes parallel to y-axis. The reason being that the slope m of such asym- ptotes is infinite. Now we shall determine asymptotes parallel to y-axis. The gencral equation of the curve of nth degree {equation (1) of art. 10°3} arranged in descending powers of y is yd (+p fy (x) Ay" o2 (X)+...=0 (1) where (x), $1 (x), $2 (x) ete. are polynomials in x. Dividing (1) by »", we get $ (X)+A/y) 61 )+A/y") $2 (x) +- Now as y->00, let xk. : From equation (2), we get $(k)=0 Thus k is a root of ¢ (x)=0 0 (2) 256 ENGINEERING MATHEMATICS Let ¢ (x)=0 have the roots k,, kz, kz etc. Then the asymp. totes parallel to y-axis are x=ky, x=kz, x=ky ete. Thus the asymptotes parallel to y-axis ar¢ obtained by equating to cero the cocficient of highest power of v {if nota constant) in the eguation of the curvy. However, if the coefficient of liighest power of y isa constant or has imaginary factors, no asymptotes parallel to yeaxis exist. oo Now asymptotcs parallel to the previous methods when 17 determine these asymptotes separately. by the following rule. The asymiptotes parallel to x-axis are casily determined by equ. ating to zcro the coeflicient of highest power of x (if not a constant) jn the equatioz curve. However if the coefficient of highest power of A tor has imaginary factors, no asymptotes parailel to Exampie 1. Find the asymptotes parallel to the coordina axes of the curve, axis are easily determined by However, it is convenient to yx—a (x4-a)=O. Sol. The equation of the curve can be written as (y'—a*) x- a= 0 Here highest power of x is on So asymptotes parallel to x-axis are ) Also highest power of y is 2. and its co-efficient is x. So asym- ptote paralicl to yraxis x= 0. Hence the asymptotes parallel to co-ordinate axes are sta and x=0 EXERCISE 10 (4) Find the asymptotes, which are parallel to cither axis, of the following curves. 1 x ya (8+ y)=0 : (D.U. 1978) and its coefficient is * y* @=O0ory= +a. yw SYN 108, Important Deductions (From Art. 10°3) G) The maximum number of asymptotes of an algebraic curve of nth degree cannot excced nm. We have seen that the slopes of the asymptotcs of the curve which are not. paralle! to y-axis, are given by the roots of $,(n)-70, which is of degree # and hence can't give more than 7 values of 7 ASYMPTOTES 257 If there are asymptotes parallel to y axis, then it is obvious that degree of ¢, (m7) will be less than” by at least the same number. Gi) If some of the roots of $n(m)=0, are imaginary. the asymptotes corresponding to these roots are called imaginary asymptotcs. ( There may be cases when even corresponding to real! roots ofgata)=O, no asymptotes exist. For example, the parabola dax has no asymptotes even though ¢, (m) =0 has real roots. 109. Other Methods of Determining Asymptotes (t) The asymptotes of an al ic curve are parallel to the lines obtained by cquating to zero, the linear factors of the highes degree terms in its equation. Let (ax) be a non-repeated factor of the nth degree terms of the equation of the curve. The cquation of the curve can be written as (van) Fiat Un -1=0, “Uy where Fu-, is a polynomial of degree (m—1) and Us, contains terms of various degrees. not higher than (1-1). It is evident that one of the roots of $n (m)=0 is m=a and hence an asymptote of the curve is yra © where c is to be determined. By definition we know Lt (y—ay) woos on (1). “. From (1), we have c, where (x, 1) lic 1 Taking limits on both sides as x~oo, we have Lt (ymax Lt Un-y! Fa x00 x00 Lt Up-y! Fa x00 Hence c is determined. (2) Let the terms of the highest degree in the equation of the curve earepeated factor (axl. Then the equation of the curve can be written as (ran? Fag (vm an'n eh 0 wel where F,-2 aid Us. 2 contain terms of degree Gr--2) and Vs egn: tains terms ofall degrees none of which is higher tian (1-21, nl t obvious that ¢,, (m1) O has a repeated root a. 258 ENGINEERING MATHEMATICS Hence there are asymptotes of the curve yay rax=c. The values of c; and cp are determined by the method given above, using the fact Lt_ yix=a x00 The following examples illustrate the use of these methods. Example 1. Find the asymptotes of the curve (82 7)32— 4x7) OB TINY AIX? — 2V9— x? 3xY—1=0, So}. The curve has no asymyptotes paralle! to the coordinate axes, asthe co-cfficients of highest powers of x and jy are both constants. Factorizing the highest degree terms, we get (x= PEP) 2x) (¥ 2)— OXF SPY By? — 2 22-4 Bay— =0 Hence the curve has asymptotes parallel to the lines x—y=0, xty=0 yr2x=0 and y+2x=0. Now from (1), we have _ 2yP4+x?—3xy+1 2 axtixy 6—5(y'X)—3(y/x)2 4-20 y/ x84 W/x— 3(y/x) - Wet 1? - CF yiIOTE— 2GTEF 2) [Dividing N* and D* by x] Now taking limits on both sides as x-+ 00, then Lt Q—x)=c x00 or Lt (x-y)=-e 200 and Lt y/x=1, we get (2 Slope of line is 1) Lx ___ 6—5~342 T= FNC —2)U-+2) oe c=0. Hence the asymptote corresponding to the factor x—y is 2 or y=x nilarly we can find the value of c corresponding to other ct These values of ¢ are —], 0 and —1 corresponding to the factors (y-t.x), »—2x and y+2x respectively. jsvMPTOTES 259 ence the asymptotes are a yYtx+1=0, yo2x and y+2x-+-l= Example 2. Find the asymptotes of (x—y)(x— 29) +7 — 3xy+2y2—7=0 (1) Sol. There are no asymptotes parallel to the axes as the co- efficients of highest powers of x and y are constant. The equation of the curve may be written as (x—yMx— 29)? + (x— yp) x—-2y) -7=0 o-(2) The curve has asymptotes parallel to the lines x—y=0 and x-2p=0. As (x—2y) is a Tepeated factor. there will be two asym- ptotes parallel to the line x—2, Now + G3) [Dividing N’ and D* by x*] Now taking limits as x00 then Lt (y—x)=c x00 or Lt (x-yl=—e x00 and Lt »/x=1, we get x00 y—x=0 se y=x is an asymptote. Now the curve has two asymptotes parallel to the line x—2y=0 We have to find Lt (y—hx)=c x00 or Lt (x—2y)=--2e x-¥00 Dividing (2), by (x—y), we have ’ —2y)—-— 2 (x—2y)*+(x—2y) x —yiy = Taking limits as x ©o, (x—2y)>—2e and y/x=4 260 ENGINEERING MATHEMATicg Hence the asymptotes are y=ixt0 and yHir+h. or xee2y or x—2y+1=0. Thus the asymptotes are y=Ex, x=23, x—2yt1—0. Note. It may be noted that the asymptotcs of this curve ip these examples could also be found by the previous method of Art. 10°3 or 10°6 10°10. Asymptotes by Inspection If the equation of a curve of wth degree can be put in the form tab tn-2=0 where un_z is a polynomial of degree not higher than (#—2) then every linear factor of wn, when equated to zero will give an asymp- tote, provided that no two straight lincs so obtained are parallel or coincident. Example 1, Find the asymptotes of x 2 wea Sol. Here the equation of the curve can be written as Untun—2=0 x) where Un =F . Upp. The asymptotes are given by un= and ASYMrAvins eur Example 2. Find the asymptotes of the curve V-6xY +x y—68+y+x=0. Sol. The given equation is of the form tn+un-2=0 Hence asymptotes are given by V—6xy? + 11x*y—623= or (y—x)(y—2x)(y—-3x) = 0 or yex, ypH=2x, and yp=3x. 1011, Asymptotes of Polar Curves Before proceeding with the determination of asymptotes of polar curves, we shall obtain the equation of a’straight line in polar form. Theorem 1. The polar equation of a straight line is p=r cos (8—«) where p is the perpendicular distance from the pole to the 2 the angle which this perpendicular makes with the initial We know that the equation ofa straight line can be written as x cos &-+y sin o=p (1) where p is the length of the perpendicular from origin (pole) on the line and « the angle which this perpendicular makes with x-axis (initia? line). Putting x=r cos 9, y=r sin @ in (1) we have the polar fom of the line as r (cos 4 cos «-+sin @ sin «)=p r cos (0—«)=p or p=r cos (6—a) +2) 262 ENGINEERING MATHEMATICS Theorem 2. The line r sin (@—y)= is an asymptote of the curve 1 CO 1 t=f where y is a root of the equation /(#)=0. Let P(r, €) be any point on the curve and O/.a_ perpendicular on the line (2). Further let Pf and PN be perpendiculars on the line (2) and OL respectively. Now PM=LN=OL—ON or PM=p-r cos (6—«) +3) Now roo as the point moves to infinity along the curve. Let 6>2, when r->co Now if line (2) is to be asymptote to the curve 1 IO) then PM—>O Hence from (3), we have as roo. Lt cos (@—«)= oy - e=y— o z This gives the value of «. Again p= OL is the polar subtangent to the curve at infinity, ie. at the point where the asymptote touches the curve . -[p: #2 * nl? Fao A) Let - w= tay : AE At i de de Thus from (4), we get -[- do P du Jato=y Therefore the equation of asymptote is — a8 = 1+ [ du jato=y r cos (4 ce 1 « or Fm sin G—o) a ASYMPTOTES . . 263 Hence the equation of asymptote is : 1 6—-y=—L.- r sin (8—-y) F@ Example 1. Find the asymptotes of the curve r=a ran 6 (DU. 1 roa 2, 76, 1975) Sol. Therefore a cot 6=fi6) (say) oo Now fi) =0 ) when cot 8=0 or G=nn+ = (say) Thus from (1), we have FO =— + cosec? @ : Feds [eS coseet * a Now the asymptotes are r sin (@—Y)= Far = ‘ when Y= the asymptote is rsin ( e-= )=-4 or rcos §=a when Y= the asymptote is r sin ( 0-35 )=-< or rcos @ Now when 0= 5, the asymptote is rsin ( 0-43) ——a or rcos 6=a, which is same as corresponding to 2 Hence there are two asymptotes only rcos @=a and r cos @=—a, Example 2. Find the asymptotes of the curve 18 =a. 264 Sol. Here the curve is r=a/é tat 6 a r a 1 Lg gre o +7? “f(é) Now S()=9 when 6=0=¥7 * 1 LOG roa or 2 Equation of asymptote is i = r sin (8—Y)= F7q@y or r sin @=a. “ ENGINEERING MATyp,, THEMA Ty, ICs “(y EXERCISE 10 (c) Find the asymptotes of the following curves : 1. rsin20=a cos 38 2, rsinnO=a 3. r=asec 0-+d tan 6 4. r=a log 6 =_4_, — 3a sin 6 cos 6 5. 1=F=cos 6 6, '= Sin’ 8+ cos? o Yayye as 3 a z & 2 a 4 Exercise 9 (c) (Page 233) a a $ 2 3 2 5 V2ar 4. abt PP Exercise 9 (d) (Page 240) iL, 16 = ( 7 2. (—2,3) fetta, — ao) o} fa+3ut/9 pe, bse py {-sar2o-+aty, 0+ (S84 )} {2. = ee Exercise 9 (e) (Page 246) 2+V2> 3B+2V2 ,—-Qt+v2) {54 Ae —V2 3— —@-v2 and (23 ae eed ©, 0), (S- n—3v3 ) for concavity 1/2 —_ (ayh- 6" /) vintervar SE to oh convex outside this interval. Exercise 10 (a) (Page 255) Y+x=0, y=O, x+1=1 YHX+1, YZ =2x, y—3x=2 =x, yet], y+X=O PEE Vtxt) 0, 2y-tx=1 o: baa, ytx=l yoxt Jao, yo3x+3—0, 2y+x+3—=0 y=0, y+x=0 8. yO, x= 1, yox+2 506 . ENGINEERING MATHEMATICS 9. y+x=41, y=0 10. yo x. Exercise 10 (b) (Page 256) ld. xetVe,yetva 2. x=cha 3. y=t1 4. x=+1,y=t1 5. x=cha, y= bb. Exercise 10 (c) (Page 264) 1. 2rsin @ =a, 28 2. rsin ( 0—-4E )= 4 where « 18 an integer. nm) weosk® 3. rcos 8t+b=a 4, 8=0 . x\_ 2a 5. rsin (0-= =Ws 6. atv2rsin( 64 =o. Exercise 1i (a) (Page 276) x Infinite Series 21. Definitions A series containing an infinite number of terms 1g. te. trees that occur according to some definite law. is called infinite series and is denoted by ties or Sut or simply Sun. n= The sum of the first # terms of the series is denoted by Ss. . So Suyck iy by bebo. Convergence and Divergence of Series Consider an infinite series ty tae ty Dune ty tte shee bth and let Sn uy-bus byt ei un, The given scrics Eu, is said to be convergent if Li Sw ne finite quantity and the given scrics is said 10 be divergent if Lt Ss ime tends to infinity However i Lt Sn docs not tend to a definite limit the given nwo series is known as oscillatory, series. sometimes called non-convergent Example 1. Discuss the convergence of the seri I rp ate (Sum of G.P! Now Lt Sn nop ENGINEERING MATHEMATICS 50 Since Lt Sn is a finite quantity, the given series is nto convergent. Example 2. Investigate the series for convergence or divergence, LF243 4... oe Sol, Here Sp=1+24+3+--4+0 = A(n+1) 2 (Sum of first # natural numbers) . Lt Sp= Lt 2+) 05 7 n>0 nte 2 Since Lt S, is infinite, the given scries is divergent. ne Example 3. Discuss the nature of the series ImI+1—I14-- Sol. Obviously Sn is 1 or 0 according as » is odd or even, Therefore, S, cannot tend to a definite limit, hence the given scries is oscillatory. EXERCISE 2 (a) Discuss the nature of the following series. 11 1 lt3e+ gt. 2. F4224+3% howe 3. 2 F223 4. 24+6+18+- 5. 34+54+7+ 6. k—-k+k—-kte 22. Fundamental Theorem on Infinite Series The convergence or divergence of an infinite scrics remains unaffected by removal or addition of a finite number of terms. Since the sum of a fmite number of terms is finite, so their temoval from the scrics or their addition to the series will not change the convergence or divergence of the series. 2.3. Convergence of a Geometric Series Show that the geometric series atartar®+... bar"... (i) Converges if |r | <1 ii) Diverges ifr >1 (ii) Oscillates if r <-1 ae INFINITE SERIES 51 Let Sn=a@+art*+--.--ar™* wee) Case I. When Ir <1. Now Sao 20a") = Lt, Ses ut ( Since | r | <1, therefore, Ltr o Lt Sn (@ finite quantity) ne Since Lt Sn isa finite quantity, the given series converges n-e2 if tri 1 We first consider the case when r>1 —atrn—1) Here Sn r= ar” 2 RL Se te (Se Sincer>1, Lt roo nye - Li Spree nt Since Lt Sn tends to infinity the given ne series diverges if r>1. When r=1, we have from (1), Su=atata+...n times =na . Lt Sa= Lt na+o. n2c0 7" neo Hence the given series is divergent when r= Thus the given series diverges if r>1, Case TI. Whenr<—l. We first consider the case whenr=—lI, we have the given series from (1), Sa=a—a+a-—...n times Obviously S;, oscillates between a and 0, according as n is odd or even, 52 ENGINEERING MATHEMATICS When r<—1, let r=—X (where K>1) ” s,— aor) a= (= Ky") Tr 1+K allt (—1)"? 7] TK : Lt So= Lt 1s x U+(—1)™! K"] which oscillates bet- n>co 00 ween—co and +9, according as n is even or odd. Therefore, the given series oscillates when rS—1 2'4. Convergence or Divergence of p-Series Show that the series rats tot dete th tes (i) Converges if p>1 Gi) Diverges if p<1 (i) Case I. When p>1 Let the terms of the given series be grouped in such a manner that first, second, third.. groups contain 1 term, 4 terms, 8 terms;.--respectively. Thus the given series may be written as 1 141 1 1 1 1 sp t(getge )+(grtd tet qe )t~ 1 8 tra) ck <#-(+ and so on. Adding corresponding sides, we get Ji4(L4 2) a4] 1 1\ et(at a )t(gtdet Ftd } 1 | 1 — 2(n- +( ge tget gis tect (3)y"44y"" 1 \ero-a) +(4 shen or BrP 14 GPE (AR GED, INFINITE SERIES 53 The series on the right hand side is a geometric series with common ratio ($)°-?<1 (when p>1), hence convergent. * Thus the given series is convergent when p>1. Gi) Gase II. When p=1 The given series becomes, wd tates +45 4$t-- a4 34($45 r)+Gre +44 )+.. re fa14 Now Similarly 3+45 Adding corresponding sides, we get i4 2444444 444444 astatats tetra ts + Vgtya +e. 1+ z+ z +5 +. The series on the right hand side (ignoring the first term) is a - geometric series with common ratio unity hence divergent. Thus the given series is divergent when p=1. Case II. When p<1. doy 1 > 7 nm w= vy] > >> and so on. arr ar Adding corresponding sides, we get 1,1 ,1 1,1 1,11 qe tort get ge te l4+ ga 454+ gt-- The series on right hand side is divergent (by case II) Thus the given series is divergent when p co Now Un=Sn—Sn-1 o Lt un= Lt (Sa—Sn-1) noo azo = Lt Sp— Lt Sara no n>00 =K—K=0 or Lt un=0 m0 Hence the condition. Note 1. It must be noted carefully that converse of the above theorem is not true, i.e. even if Lt un=0, the series 2, may not n00 converge. For example, Iet Bun =1+ +i+ tet ta... = Here un= 5 n Lt m= Lt +0 m->cO nao 7 But we have scen that the given series is divergent. INFINITE SERIES ss Note 2. The result can effectively be applied to show that the given series is divergent if Lt ua%0. ° ne For example, let Sua=1-+2+3-+---+n+--+ Here ian o Lt wa= Lt_n¥0 nro nc Hence the given series is divergent. Tests for Convergence and Divergence of a Series Comparison Test Wu, and 5 be two positive term series, such that and after some particular term Lt 2 =& (a non-zero finite quan- a 4 and E vq either both converge or both diverge. Proof. Let the scries from and after the particular term be. tybusbigt tint. Now for all values of 1, from tity), then let Lt =k n>co Vn _ Therefore by definition of limit, there exists a posi s, such that I< efor allan ie. or k—e (ke) Yn or Eun > (ke) Eva Each term of Zu, is term by term greater than the corres- jonding terms of (k—e) Zvn, Since Zya is divergent, therefore, un is also divergent. Note 1. This test is very useful when degree of 7, in un can readily be determined. Note 2. To select auxiliary series Zye=E 1;, it should be noted that p fference in degree of » in denominator and numerator of un. Example 1. Test for convergence of the series, T,i,] ztst Sol. Here way ie Let 2va=E 1), there- fore, Zun is also convergent. Example 2. Test the convergence or divergence of the series, Bt g+h+gt..0 n+l nm Sol. Here un= fe 1 “. Now let Lm=z Gear be the auxiliary series, INFINITE SERIES 57 Yn = pri WEL oe = - nr n+l 1 =tt1 -( ~ +) . Lt we Lt ( 14+) noo 7 =1 [a non-zero finite quantity] Hence = un and =vn converge or diverge together. But Eyed Li is, 7 () Convergent ifp—l1>1 tep> 2 Gi) Divergent ifp—-1@ 1 iep <2 un converges if p > 2 and diverges if p < 2 Example 3. Discuss the nature of the series, Sol. Let Now os It “= Lt —a nc va nro (s+) 2 . - . =s [a non-zero finite quantity] Hence Eu, and Eva converge or diverge together. 1, But Evn—E—- is known to be divergent, therefore, the given series is also divergent. wn Ss ENGINEERING MATHEMATICS 2.7, Comparison of Ratios If Sun and Brn be two positive term series such that from and after some particular term, un ¥ tne, Vata and if Eva is convergent, then =v is also convergent. Proof. Let the two series uv, and =vn, from and after the particular term be, tytie tis band ytretvst---, respectively. . tn Va Since tints > Vata 1 | te , Hs! and soo Uz > vs Uy Vg uy eye ® nm Uy V3 Ug Va yee, ee 0 or uy Ye tz va Us 1. Proof. Lét the scries from and after the particular term be Case I. ty-P italy bee btn bo When k < 1. By definition of a limit, a positive number AK <% < 1) can be found such that =) Thus «and so on. + @) Now mt ratia te bitn bee Dna eth eth a th eu EAE MER) ~ By @) = 5 a finite quantity. Hence = un is convergent. Case Il. ‘When k>1. By definition of limit, we have Now Unser Un Pl. Spay titan =u(i+ e+ A ee Ee ts . Meg Ma Mag pe w(t Ge +a: Get st So> (+11 -+.--n terms) Sa>niy Now Lt S,> Lt mm which tends to infinity. +00 no Hence Zne is divergent. 60 ENGINEERING MATHEMATICS Note. If Lt AS =1, the test fails to give any information no about convergence or divergence of the scrics and therefore further tests are necded. For example Ict =144,144,1, i Buns Lb bg beg beh ote Here Un “ea=GET) Ung, n 1 ta ML (+1 /n) 1 1 = net tn nite UF IT)? The series Zun is known to be divergent, being p-series with 1. Further, let Zuw=1-++ watt oo Here w= Fe and tn GEE . Unit ne 1 ” an GF)? FI jny® b Lt tH ay nao tn But the given series is convergent, being p-series with p> 1. Thus we sec that when Lt “unt no tn =1, D’ Alembert’s ratio test fails. Example 1. Test for convergence the series, 1 2 3 Teo tee + eet Sol. Here t= Po n41 ‘ wen TET nts nL 142" To INFI NITE SERIES 61 Hence the given series is convergent. Exemple 2. Test for coivergence the series J4 BxtSx24 784... (x>0) Sol. Here tin = (2 = 1)" tne (2n+ 1X" = Qn+) Qu=T) Lt ‘The given series converges if ¥<1 and diverges for x>1. Ifx=1, the test fails and we apply further tests. When x=1, we have w= (2— 1) Le un= Lt (Qn—1)*0° ne pa Since Lt wn30, the given scries diverges for x=1. no Hence the given series converges for x<1 and diverges for x>l. Example 3. Discuss the convergence and divergence of the series 62 BINSUEN BERENS DUATHEMATICS Gi) W(t )—-¥n) sen, J (7) = sin ra Sol. (i) Here w=(i-+-1)'8—n4 sia LY? aa ams (142) on 13 Lys “(ty =n 1,1,4G@-),1 }- = [{+3 nto aT ne I 1 1 =n'8 | 7 —jate ” ( 3n ont ) =pr(+ J. ) = pela-orte Let = ae - Applying the comparison test, we have un 1 Baar ae 1 1 ~ (fsb Jeb which is a non-zero finite quantity. and 2y. either Therefore, both the scrics Eun converge or diverge together. is known to be divergent But 2rn(p=% <1) As such 2x, is also divergent. (ii) Here un=si Lee mob n Now u “es Lt sin — n+a vn ot ( ale) aot -22 oy x0 *« where x, no INFINITE SERIES 63 But 2y, is a divergent scries (p=1). Therefore Zun=2 sin —L is also a divergent series. EXERCISE 2 (4) Test for convergence of the series. ® bhp tata — Rie cae te 12. p3x FSX FIX... 1% p28 FAT s+ ( 7 (a) 5 ‘ x N ) ee ta converge when x <1 and diverge when xP 1. 2:9, Raabe's Test. [Higher Ratio Test} : If Xun be a’ positive term scrics,,such that from and after some particular term * --1 \=«, then Zum, NO Lt a(< novo \ Matt (i) Converges if k>1 (ii) Diverges ifk1 Let p be a_positive number such that k>p>1 and compure the given Series vn with an auxiliary series. which is convergent when p>1. Now Sw, converges if, (See Art. 2°7) Ly Lee] Dye ie. if or if ie. if &k>p. which is true. Le an (Geo = J Hence the given series Sun is convergent. Case II. When k<1. sitivé number such that A 1, he. 2> x or x<2 3 x The test fails at and diverges if 2. When wary Gel) on siet Jon SERB 1] 68 ENGINEERING MATHEMATIC ‘Thus the series diverges when x=2. Hence tbe given series converges when x < 2 and diverges fo, xp2- 2:10, Logarithmic Test If Zun be a positive term series, such that from and after some particular term, Lt nlog 2 n+0 uney =k, then Zun (® converges if k > 1 (ii) diverges if k 1. Let p be a positive number su ‘the given series with an auxiliary sesinghat k>p> and compare B.sE1 , ne “which is convergent when p > ], Now Zun is convergent if un va Yen any or if —¥e_ L (n4+1) tn > a (ye nu or if log 2 Ung > P lo ( 4+) na UF 1 ootae (sa LY if a ri eee > e(ho n ~ Daa te.. ) or if nog — Me tg > P~ By or if Lt Un it, log INFINITE SERIES 69 or if k > p which is true. Hence the given series is convergent. Case II, The case when &k < 1 can be proved similary- Note 1. The above test fails when k=1. Example 1. Test the convergence of the series (a—19"7?- SSS @—p Sol. Here w= GOD"? Una ey us (a1? | (nt o: un nm oan wet (1-2 YO (eh ne AE -Gety oe wa (FY (04 EY (144) . nlog gaan [e—» tog (1-4 ) +108 ( itt) siog (14+ J -n[a-y(-4 +--+.) tat a(E dee J] ofthe] ENGINEERING MATHE MAT, MN +1og ( 4+t)e(4 rt... )] =. o[( 1437...) n> Hence the given series converges, Example 2. Test the convergence of the series 4 Sol. Here tins 1) xn and dan GDS unger — (nt1)" _ x an un nek atl in py_od = Thus Lt, i ae Lt (145) rex jn ae sect . ; herefore, the serics is convergent when ex <1 and divergen| when St ‘When ex=1 or x=I/e we apply further test. une nad } mies as = Li mles Corer: eet = Ltn log (cy) - n = Ltn {ice e—(a—1) tos ( et I j ee eel {FINITE SERIES ~ efininn (dghesbon =. {»-c-n+@5D ee} awe aA = 31 _ @=1) }-3> i fhe, As such the given series is convergent when x—1/e. EXERCISE 2 (c) Test the following series for convergence or divergence. 1 1 (*>0) (e>0) (x>0) 2 a+ 1)P(B+1) HC 1)(@+2)8(B+ 1)(B+2) 6 AEE xt OEE 1.2.3 YOr-+ 1G 2) e+ 211, Cauchy Root Test Wwe un be a positive term series, such that from and after Some particular term, Lt (up) ke, then Eun nc (@) converges if k <1 (Ui) diverges if k > 1. Proof. Let the series from and after the particular term be " yb bbe bunb Case. 1 hen k <1. such BY, defini on of a limit a positive (k < I <1) can be found ‘Uch that ” (uy 1. By definition of limit or (unyle>1 ie. my > 1, ue > 1, ty > Lye ee DL we uy tttg big bo eee tab vee > PPIAI oe 14 oe The series on the right hand sides is obviously divergent. Therefore the given series Sun is also divergent. Note. The test fails if k=1. Example 1, Discuss the convergence of the series, Jat ta Itty tent ate = i tI ce iL Sol. Here Un? oe or (tn) a =0 <1 ale ve Lt (un)'"= Lt nto a> Hence the given series is convergent. Example 2. Discuss the convergence of the series. SCs4)" Sol. Here . un=( 144+)" . iy *% ayn + (un) ( 145, a Lt (un)'"= Lt ( 141 )'=e >1 n>> ne n Therefore the given series js divergent. Example 3. Test the series for convergence and divergence, f.xt(F) 2t(Z) ett (Sty yh ote yee INFINITE SERIES 73 Sol. = nr ol. Here un (SE ” WET * =(—1_y wale ) Lx = him o Lt (un)'™= Lt no neo OF1/n)* The given series converges if = lv ie. x>e. The above test fails at x=e. When x=e u =(s a NT “¢ Now Lt wn0, therefore, the given serics diverges at x=e. pay Hence the given series converges for x 0] For example 1—4+4—4-+.-- is an alternating series, Leibnit’z Test for Convergence of Alternating Serics An alternating series yup y= Wate [un>0] is convergent, if for all values of-n, @) van Untry uy >> us> -.- <= (2) From (1) and (2) it “follows that the expression in each of the brackets in (1) is positive. Hence the sum S¢q is positive, ie. Sin >O Now =U — [tg 10g) + (tgs) +--+ (ten — onta)] The expression in cach of the parenthesis is positive’ and subtracted from 14. oO Sontr Quy Also Sonti=San tanta oc Lt Senty= Lt Sent Lt ugar +3) +0 nto neo The given series is convergent if Lt Sints> Lt Sun +(4) n> nso Therefore, from (3) and (4), we have Lt uanty=0 for alien, no Replacing (2n+1) by 1, we have or Lt un=0 ate INFINITE SERIES 75 Thus the given series is convergent if G@) mou >ua>... Gi) Lt un=0 ne Note 1. It must be noted carefully that both the conditions are satisfied simultancously for convergence of an alternating scrics. Note 2, An alternating scries if not convergent is oscillatory. Example 1. Discuss the convergence of the series, I i w-$4+4—f+.. Sol. The terms of the series are alternately positive and negative, and also each term is numerically less than the preceding term, ; ie. n> tines for alla (1) Also Lt m= Lt +=o o(2) a> nao na From the conditions (1) and (2), the given serics converges. Example 2. Test the convergence of the series, Z Fete i ZT I—# te Sol. The terms of the scrics are alternately positive and negative and also cach term is numerically less than the preceding term, i.e., Un>Untr forall a -G@) 1 Also as o nt, Gna « Q) From the conditions (1) and (2), the given series converges. Example 3. Test the convergence of the series, i et xt x- Rta Oe Sol. Here we shall apply D’ Alembert test, x 3 n= (1) nth and unta=(— 1)" Gye 16 ENGINEERING MATHEMATICS tnd, xn n a a = Pa |= esto * . Lt jess pee | un 11 G+ime '*! =Ix/ ‘Therefore the given series converges, when |x| 1. When | x | =1, D’ Alembert’s ratio test fails. Now when | x | =1, the series can be shown to be convergent ‘by Leibnitz’s test. Hence the given series is convergent for | x | <1 and diver- gent for |x| >I. EXERCISE 2 (e) Examine each of the following series for convergence or divergence : a 1-d4+4-$4+-. a E-detd ed be 3 3-4 4. s Sor 6. oe need 7 lag the te 1 1 1 _ 1 & yor visi yal vast 2:13. Absolute Convergence Aseries 2un is said to be absolutely convergent if =u, and 2] unl both converge. INFINITE SERIES 77 For example let Here Eu, is convergent by Leibnitz’s rule. Also Vin tt t+ Jet..co is convergent. being a geo ric series with commen ratio 4 (<1) Hence the given scrics is absolutely convergent. 2:14, Conditional Convergence A series Sup is said to be conditionally convergent if Sun con- verges but | uw, | does not converge. For example let Zun= 1 1 i ytpo gt Here Eu, is convergent by Leibnitz’s rule Now Eun t s1t$4+$4+t4..0, is a divergent series being a p-scrics with p=1. Thus =u, converges but © | un | diverges. Hence the given series =u, is conditionally convergent. EXERCISE 2 (f) Find out which of the following series converge absolutely or conditionally : 1 1 fog? fog fog t v2 . (= ‘ ane 6 Veep - w+) 1 Yi. % loge tes ya te ~ ANSWERS 495 Exercise i (c) (Page 34—35) 4. @) cos « cosh 8—i sin « sinh B Gi) sin 22 sinh 28 cosh 28—cos 2%" cosh 2B—cos 22 sin « cosh B . cos « sinh B ] (ii) 2 [ets 2B—cos 2a’ cosh 2B—cos 20, @) cosh « cos B-F/ sinh « sinh 8 ©) “sinh 2« |, ___sin 28 cosh 2a-+cos 2p '* cosh 224-cos 2p Gi) eCOsh x €os » [cos (sinh x sin ¥)-Fi sin (sinh x sin y)) Exercise 1 (f) (Page 42—43) 1. (i) cos7(/ Sin 6)+i log {V sin 0+~/I-Fsin 0} Gi) sin“ (W Sin 6)-+/ log (V1-+sin 0—4sin 6} iy 2 : 2x i tan { —22 Git) J tanh ata }+ pian fe 12. e27™+= [cos (log sec #) —i sin (log sec «)] 13. Exp (— 3) cos (+ Jog 2) Exercise 1 (g) (Page 48) 4, 4sine 2 Wlaxcosy * 54 cose | ToIxcosyty 3. cota 4. excosB . cos (2-+x sin B)- 5. 78 sin (sin 8). 6. } tan ess , (#1, 0% nn)- kad tr . : ses ive. 7. ar a according as cos « is positive or negative. 8. (2sin $y" cos eH) 9, sin b4-(—1) ant (x sin n6-tsin (+1) 9 . x'+2x cos +1 10. 1 +( 2cos > $y: sin a : Exercise 2 (a) “Qose 50) 1, Convergent 2. Divergent 3.. Divergent 4. Divergent 5. Divergent ‘6. Oscillatory Exercise 2 (4) Page (63—64) 1. Divegt. 2. Divet. 496 avayne ENGINEERING MATHEMATICS Convgt. 4. Convet. Convegt. 6. Convet. Convst. 8. Convgt. Convet. 10. Convet. for x<1 ; Divgt. for x>1 Convat. for x<1 5 Divet. for x>1. Convgt: for x<1 ; Divgt. for x>1. Convat. for x<1 } Divgt. for x31. Convgt. 15. Convat. Convet. 17. Convgt: Convgt: 19. Divgt. Congvt. 21. Divat. Convegt if p>2 and Divet. for p<2. Convat. 24. Divergent. Divet. 26. Convet. Divgt. 28. Convgt. Convst. Exercise 2. (c) (Page 71) *Convet. for x<2 and Divgt. for x>2 Convat. for x<# and Divat. for x> Convet. for x<1 and Divgt. for x>1 Convat. for x<1 and Divgt. for x>1 Conver. Convat for x<1, Divgt. for x1; when x=1, Convgt. if --—a—@>0 and Divgt. for + BO, Exercise 2 (d) (Page 73) Convet. 2, Convet. Convet. ; Convgt. for x<1 and Divgt.forx>1. 5. Divgt. Exereise 2 (¢) (Page 76) Convet. 2, Convet. Convgt. 4. Convet. Convat. 6. Convet. Convegt. for | x | <1 and Divgt. for | x | >... Divat. Exercise 2 ({) (Page 77) Abs. convat. 2. Conditionally convet. Abs. -onset. 4, Conditionally convgt. mally convgt. Conditionally convgt. 6 Condi Conditionally convgt. Exercise 3 (Page 84—85) 4. 56-1) ~I5GF4) 1 ~3fs—=2)~ Expansions of Functions and Indeterminate Forms G1. The student is already familia: with expansions of ele- mentary functions using Binomial Theorem. In this chapter we shall expand the given function as an infinite convergent scries in the form dptayx+a-x*-+...+a,x"+--., known as power series. It is assumed that all the functions dealt here possess finite and con- tinuows derivatives of ail orders for the values of variables under consideration and are capable of expansions as power series. 62. Maclaurin's Theorem If a function f(x) cam be expanded as an infinite convergent series of positive integral power of x, then SOP ONV+AS OF FO) on A $000) where f"(0) stands for nth derivative of f(x) at x= Proof. Since f(x) is capable of being expanded as an infinite series, let SC) Sag bayxbayt age tat +... =) By successive differentiation, we get SCO a, 42.02x43.0:x"4 dae... SC) 2.43.2. 24+-4.3.04x +o. SOC) = 3.2.05 4-4.3.2.04x Substituting x=0, successively in (1), (2), (3) and (4), we get fO)=ay or ao—=f(0) S'O=a, or a=7'0) S'(O)=2.49 or Lo ss SO)=3.2.a, or a,= FG ana so on ‘Substituting the values of ao, 41, @2, as, etc. in (1), we get Rey xan x” SISO +S O)+ 77 F O+35 Ff (0)+---+ atl" O)+, The series on the R.H.S. is known as Maclaurin’s Series. Note 1. Another useful form of the above series for the function »=f{(x) is, I=Y=W)oFODE+ 0 Fp 7 +Ono aT oboe where (yn) Stands for the mth derivative of y at x20. 63. Expansion of sin x Let f(x)=sin x “. {(0)=0 Sf'(x) =cos x S'O)=1 S'(xda—sin x f'(0)=0 S''(x)=—Cos x tO. f'*(0)=sin x S**(0)=0 and so on. The values of derivatives atx=O are repeated in cycles of 0, 1, 0, = By Malcaurin’s Theorem, we wave fe) =f0) +f 0) + FS ro+k SrrO+e, FO) +. “Sin x= x a —#+e Aliter Here f(x)=sin x - nt . f(x)=sin ( at) Putting x=0 on both sides, we have S(O)esin Substituting n=0, 1, 2, 3, ..., we get FO) , A, ’ g ‘@=sin zo S'(0)=sin ==0 F"Oesin FF = J'*(0)=sin 2x=0 and so on. Hence by Maclaurin’ 's Theorem, we have x E xé sin x=x— atts Fite EXPANSIONS OF FUNCTIONS AND INDETERMINATE FORMS 135 64, Expansion of a* Let f(x) =a? 2. (0) =a°=1 S ()= a7 loga J'O)=log a SCX) = a? (log a)? £'(0)=(og a)? £10) =(og a)* S'"() =a? (log a)? Proceeding in this manner, we have, f"(0)=(log a)" By Maclaurin’s theorem, we have LDAOS OFS OFF SO + Ty at =1 F(x log a) +54 Ge log ate. (log ob tgp Glog a+. F"O) +... Note. The expansion of e* can be obtained by putting a=c, in the above result so that log a=log e= . etelaxt Seb ee... 7 2! 31 = nt 65, Expansion of log (1-+x) Let f(x)=log (1+ >) ©. CO)=log 1=0 oda “(0)= SO=TEE F'O)=1 “=, “(= — S°O= Ga F°O)=—1 £7@= GREP £7"@)=21 Siete WEARS) p09) = —31 and so on. By Maclaurin’s Theorem, we have WSOFS OB SOF Fy srO+ A FO+-- 2 log Gta—x4+ 2) n+ > oy xt xt tT aot =x— J Expansion of log (1—x) can be obtained by replacing Q)+ aT {-@G)}+..- Note. x by (=x), in the above result, “log U—x).= 136 ENGINEERING MATHEMATIOg Example 1. Show that 2 2 sins? x=x+ Het F a z x44 Ot . Hence find the value of ™. Sol. Let y=sin"?x 4 Differentiating both sides w.r.t. x, “(1) n=T a) or yi (l—x)= Differentiating again W.T.L. x, we get 2ysy1—39)—2y2.x=0" or yx(1—x*)—yyx=0 (3) Differentiating both sides of (3), n times using Leibnite; Theorem. . Dnt) 47°C Yat (—29)+7C2. Pal —2)] —Tynta x+°Cs-yn-l]=0 or x") ynt2— (20+ 1) x¥nta— nin =0 wwl(4) Putting x=0 in (1), (2), (3) and (4) we get O=0, o=1, (¥2)o=0 a and (nt2)0—77(Yno=0 or (Ont)0=n*Yn)o (0 Substituting n=1, 2, 3, 4, 5, etc. in (6), we get On=17 (i)ysl? ty Oden Odo=2.0)0— Lv Odert On)o=5*.(ys)o= 57.38.12, By Maclaurin’s Theorem, xt robes a dot Over F am +3 “ot é Eva Fr x ya sin? x= Otx1+ F. o+ = res o+ a7! pt 2385 a oF Fy" or sin mat BE gay IE gy VE atte EXPANSIONS OF FUNCTIONS AND INDETERMINATE FORMS 137 Putting s=24 on both sides, we get wt Dota l(1V, 3 (Ly sin™ 2aatela) tale) t- = a 0°5000-+-0°0208+-0°0023=0°5231 =3'1386—3°14 (approximately) Example 2. Expand tan ($+ jin ascending powers of x. Hence find the value of tan 45° 30' to four places of decimals. Sol. Let Sed=tan( zy x) £(0)=tan (=)=: S'@=set (F+x ) o f@)=sec? F—2 f°) =2 sec (= +x ) tan (=+* ) =2 {tant (#+* t tan ( +x) =2{ tan ($+ J+ tan? ($+ J} “ S°O)=4 S)=2 { sect (3+ )+3 tant( Fx ) sec? (+= yt 2 scet( F +x yf 143 tant ($+ y} =2 {i+tant (F +x J} +3 tant (4 +x yy - —2{ 144 tan?( B4+x )+3 tan'( += )} 2 S'’'O=16 FPO)=2 { 8 tan ($+=) sec? (4+ ) +112 tan? (F +x ) sec* (3+ )} -. F0)=s80 By Maclaurin’s Theorem {9 =10) 42f'0)-+ AZ ro+ArOrg {T° O+. ENGINEERING MA; THEY, TH 138 . tan G +x )atd2et H @+ar O64 (80). =142xt 20+ = e+ 2 Xow . Putting = ¥=30°'= 30° ial x a tan 45° 30'=142- gp +? ° (360) + =1+01 745+0°00015=1 “0176 (approximately) Example 3. Expand log [i—log (I—)lin powers g Machen Mcorens upto the term of x and deduce the expo? of log (i -+log U+3))- Pn Sol Let f(x)=log [I —log (1~log 29] ‘We know, log (1—x)= 2 2 sad-tos[ 1-( FF —~ J] =log [1+( xt oe y eye Let xt ty +--=2 “ F(x) =log (1-2) =2-£42 =z2-7 tte -(s$+$e- 4 +4( a+ 4 e a FR. art Ete. log [1 log (1—x)]=2-+ +... eel) 7 x. ;: Replacing x by T+z in both sides of (1), we get log [1+log 4x) =x 4+2)744(1 +23 =x(1—xtx2—...) $41 —3xt- te" £ Vogl Hog l+x)J=x—xt4 2 4 2 . > ‘EXPANSIONS OF FUNCTIONS AND INDETERMINATE FORMS 139 Example 4. Apply Maclaurin’s Theorem to obtain the ex- pansion of the function e** sin bx in an infinite series of powers of %, giving the general term. Sol. Let f(x)=e* sin bx then SF) = (a?+52)"!7.€% sin (6x8) we) where Is S20} “ and cos 8 ae Now I7(0)= (a? +87)" sin (19) (from 1) Putting n=1, 2, 3. ... successively, we get S'O)=(e2+b?)" . sin 8 asay2 b SOT ONE. Tate? FO)=E FD. sin 20 =(a*+5?) . 2 sin 9 cos 6 2ab =(a?+b?) Gree A 208- F'' O)=(22+b2*"* . sin 36 =(a?-+57)*!? . (3 sin 6—4 sin® 6) : sey [pe ei] =b(3a?— B*) Also f(0)=0 By Maclaurin’s theorem, FOSO +S O+ Ze fot FL "O+... b(3a*—b?) AO 66. Failure of Maclaurin’s Theorem It should be clearly understood that every function cannot be expanded by Maclaurin’s Theorem. This theorem is not applicable in the following cases. () The function f(x) or any of its successive derivatives do not exist finitely at x=0. % e™ sin bx=bx +abx*+ (ii) The infinite series obtained by expansion does not con- verge. For example Maclaurin’s Theorem c; be-applied to obtain the expansion of functions like cot x, log x etc. 140 ING MATHEMATICS 67. Taylor's Theorem jon fix4-h) can be expanded as an infinite convergent integral powers of " then Wa fan scrics of posi LOAM=SOI LYON es “()+- or SX) ee LE IMa +. om where f* (x) stands for the nth derivative of f(v+-4) with respect to (x-FA), when (v+-A) is replaced by wv. Proof, Since f(.v-+ h) is capable of being expanded as an infi- nite scries in powers of /, ANE MN = ag ag} agh™ ash? agh*+ sel) Let us find derivative oF feet). Now <# S(xtA= ae + feta). ae (xh) HSf'(x thes (eth) Also ary Stas (eM Hence differentiation of f(x+h) with respect to (x+A) orh gives the same results. Differentiating (1) successively, with respect to h, we get S' (xt Sa,t2. agh+3.ash?+4.adP +o SRAM =2.ag+3.2.gh+4.3.0¢h" 2d +4.3.2 adh. , in (1), (2), (3) and (4) etc., we get we Go=S(x) a,=f"(x) oh SO) =3.2.45 an fo Substituting these values of do, @;, 42 and ay etc. in (1), we obtain LAMPDLM CI+ [OL SAL PO. The series on R.H.S. is also known as Taylor’s Series. Note J._A function f(x) may be expanded in powers of (x—. by Taylor's Theorem by putting A=x—a. p Ga 2 Sad =flayt (x alf (+ FEM poy =O Sa +on EXPANSIONS OW FUNCTIONS AND INDETERMINATE FORMS 141 Note 2. Putting X=0 and A=x in Taylor's series, we obtain Maclaurin's Series SL) SO) EAPO) + 37 roy ~f''O)A o 3 Thus Maclaurin’s Scries can be obtained as a particular case of Taylor's Series. Example tl. Expand log, (x+-h) in powers of h by Taylor's Theorem. ° Sol. By Taylor's Theorem, 2 . POEM=S COAL COE frOO+ wer 1) bene EE px) bene soa , log + log xt — fh a yt. e+... 2 Sol. Sin x may be written as sin Le+( x $-)} “Here x is 5 and his Example 2. Expand sin x in powers of ( x-3 ): Now f (x) =sin x or SF’) =cos x S'G)=—sin x S'M = —cos ¥ F*(x)=sin x By Taylor’s Theorem, + Ime sin x=14+0— soul x or sin x=1— a 44 Example 3. Prove by Taylor's Theorem. tan (x-+h)=tan x-+(h, sin @). sin a—(h sin «)?, Sit 2x 2 +(h sin «)*, Sin Be where x=cot x. Sol. Here f (x+h)=tan™ (x+-4) aA SF (x)=tan7 x 4 1 1 L'O= TEs Treo a =Ttosecra sina f'@=— 2x _- _2cote +x? (+ cot? a) 2 cote « -=—2 cot & sint « _ =— 21-3 cot? «) ~ aE (i-+Feet? a)F =—2sin? «—3 cos «) sin’ « =—2(4 sin? ¢—3) sint o =2(3 sin «—4 sin® @) sin? a =2, sin 3a. sin? « - 3 sin o—4 sin? e«=sin 3a) By Taylor’s Theorem, S (xtid=f +e wot & , sets fre )b ane EXPANSIONS OF FUNCTIONS AND INDETERMINATE FORMS 143 2. tan —* (x+A)=tan“ x+(h sin «) sin = —(i sin 0)* - ee +-(h sin 2)? . si. se Aliter F@)=tan x “ S)=(—1)"? . (1) E sin” & sin no Substituting » 2, 3+, we get ==sin &. sin & Co ol Ff" (x)= sin? « . sin 2% Ff (#)=2 1 sin? . sin 3 By Taylor's Theorem, tan” (x+-A)=tan x+(h sin <) . sine —(h sin a) . 18,2 4. (4 sim oy, 1 3E 4, Example 4. Apply Taylor’s Theorem to eetene the value of AZ) , where fa) — x94 3x74 15x—10. Sol. By Taylor's Theorem, we have S(R+FA)=S)+ Af’ w+ Tt” & Eat... Put x=1 and a > o f( 1455 Ja rsh r w+ Geyrro 4h (Gb yrvo+. eb Now S()HP4 39+ 15x—10 2 S0)=9 SF) =BHEXHIS s'y=24 F')=6x+6 SF Q= 12 S''W=6 f'')=6_ All other derivatives of f (x) vanish. Substituting these values in (1), we get ll 1 1 (43 J=94+5 2 Mtge U2)+-glgr 10° =9+2°4+0°06+0°001 =11°461 Ise ENGIN SEN ee Example 5. Given sin 30°}, use Taylor's Theorem to evaluate Sin 31° correct to four significant figures. (cos 30°=0'8660) Sol. Let f(x+h)ssin (x +) o f(x=sin x in x and so on By Taylor's Theorem, we have sin (x+A)=sin x+h cos x— & sin x—.. (1) and A=1°= radians in (1), we get Putting cn ee aw) “sing =0°5-+0'0175 x 0°866—¥ (0°0175)* X0'S—... =0°5+0°01515—0°000076—.-- =0°5151 upto four places of decimal. 68. Expansion by Differeniiation and Integration of a Jmown Series ____ These methods are useful, if the series for a given function is knowa and it is required to obtain a_ series for its derivative or integrals. The following examples illustrate the use of these methods, a x Estmple 1. Using the series for sin x, obtain the series for x xe aT tain Sol. We know sin x=x Differentiating both sides with respect to x, Example 2, Find by integration the series for. (i) log, (1+x) (ii) tan x. Oe. Sol. Now peu tx) =1l—x+ We get teerating both sides with respect to x, between limits 0 andjx, EXPANSIONS OF FUNCTIONS AND INDETERMINATE FORMS 145 x x J us dxf (—xta0— Adds 2 . or loge U4+x)=x— +--+... w Ga teyteottstiwt te. Integrating both sides with respect to x, between limits 0 and x, we get x x Jaks axl (xt xt pede oO ey xs or tant xax— $+ 6.9. Approximate Galculations Let x and y be two variables related to cach other by the relation y=f(x). It is often required to find the change Qrin y when x changes by a small amount Ax. This can be obtained by Taylor's Theorem. Now »y+Ay=f(x+ A>) Bo Ay=Slxt+ Ax)—y =S(x+ Af (2) b=fod] ” AYO AS WIAXETG SOMME IAS) [By Taylor's Theorem] Now as Axis small, its square and higher powers may be neglected. Ay=S' (%) Ax or ayn - Ox Ax If Ax is error in x, then = is called the relative error and “2 > 100 is called the percentage error. Example 1. A circular plate expands under the influence heat so that its radius increases from 5 cm. to 5°06.cm.. Find the approximate increase in the area. Sol. Let r be the radius of the circular plate and A its area, then A=nr* * a4= 4 | Ara2rr dr Here r=Som.; Ar=0°06 cm. AA=2n (5) (0°06)=0°6 2—=1°88 cm’. Example 2. What error_in common logarith will be produced by an error of 1% in the mumben Fe mmbe Sol. Let y=log yo x =log. x . logy ¢ (Base changing formula) - d) : Ay=e . Ax= +. Ax ) logie dx x by 00 Ax | login e _ 1%x0°4343 x 100 100 =0°004343. Hence error in common logarithm is 0°004343. Example 3. The quantity Qof the water flowing over mula Q=CH®", where H is the head yenotch is given by the for water and C is a constant. Find the error in Qif the error in“y is J°5 per cent. Sol.. Let error in @ be AQ. Then AQ=-22. An. or Now {Gas CH?I?, The percentage error in @ is given by A@ yiq9= #2. . 4 AB x 100=-Fe 100 cH x 77x 100 Example 4. The area of a triangle is calculated from the angles A and C and the side b. if asmall error 8A is made in measuring A, show that the percentage error in the area is about 100 8A. sin Chisin A . sin {A+C)}. Sol. Let 4 be the area of the triangle and 3A be the error in A. Then A= bec sin A 1 ya sin Csin A 2° sin (A+C) be op ce bin _b sin € sin B Cc sin B sin (180—A—C) —_4 sinc sin (A+C) since ‘EXPANSIONS OF FUNCTIONS AND INDETERMINATE FORMS 147 h=1 psinc xf cos A. sin (A+C)—cos (A+C) sin A sin? (A+C) = sin’c ___ Zsin® (AFC) BA y 199-( 48 100 Now x x100—( ge aa)x ie C.3A,, 1002 sin (A+C) 2sin* (AC) BF sin C sina =100 3A . sin C/{sin A . sin (A+C)}. EXERCISE 6 (a) 1. Apply Maclaurin's Theorem to expand (i) log sec x (ii) cos x (iii) log (1+sin x). Prove the following by Maclaurin’s Theorem. .. n(n—1) Dn 2 C+x)=ltaxt PSS x SF ) x... an eae 3. e = 1x z 3 eee sect 2 tat 4, easin xa tax¢ 4 ate) 202-ba® + Oe) at.. and hence show that i e@=1+sin 64+ in? 6, 2 S 4 Bi +H sin" o+Zy sin 6+... 5. Expand sin (7 sin? x) in ascending powers of x. 6. If y=sin log (x*+2x+1), then prove that 2 3 5 3 ye2x— x —FZ P+ StF tS te. 7. Expand log. (x+¥x7+1) up to first four terms by Maclaurin’s theorem ; by putting x—0'75 in the expansion, cal- culate the value of log, 2 to four places of decimals and find the percentage error if any. 8. Expand log, cos x by Maclaurin's theorem as far as the term x‘ and calculate logis cos 7/12 up to three places of decimal. 9. Calculate the approximate value of ¥ 10 to four places of decimal by taking the first four terms of an appropriate expansion. iHint. Expand (1+ x)! by Maclaurin’s theorem and put x=1/9.] 148 10. Expand log. sin_(x+/) in ascending powers of , . find the value of log, sin 32°to four places of decimal,” Her (Given log 2=0°69315) . Expand the following functions by Taylor’s theorem, 11. tan (x-+4A) in powers of h. 12, sin- (x+A) in powers of x. 13. log sin x in powers of x2. 14, tan7 x in powers of xy 15. Expand the polynomial f(x)=x*—22°+3x+5 in Posigy, integral powers of (x—2). itive 16. Apply Taylor's theorem to ¢alculate the value of $(2'001) if f(x) =x —2x+5. 17. Evaluate (i) sin 30° 30’ (i) sin 1° 15’. 18. Show that sin (@+4) differs from sin a+h cos q by agg 2 y hi more than zz 19. A heavy string suspended is in the form Y=a cosh Show that for small |x| the shape of the string is approxima expressed by the parabol ly - 2. yrat 57 20. Show that for |x|

You might also like