: BMTE-144 lignou Ozes NUMERICAL ANALYSIS Indira Gandhi National Open University School of Sciences Volume-I SOLUTIONS OF NON LINEAR EQUATIONS AND SYSTEMS OF LINEAR EQUATIONS Course Introduction 3 BLOCK 1 Solution of Nonlinear Equations in One Variable 5 BLOCK 2 Solution of Systems of Linear Equations 145Unitised Course Outline Volume | Solutions of Non Linear Equations and Systems of Linear Equations Block 1 Solution of Nonlinear Equations in One Variable Unit 1: Errors and Approximations Unit 2: iterative Methods for Locating a Root Unit 3: Chord Methods for Finding Roots Unit 4: Approximate Roots of Polynomials Equations Block 2 Solution of Systems of Linear Equations Unit 5: Direct Methods Unit 6: Inverse of a Square Matrix Unit 7: tterative Methods Unit 8: Eigenvalues and Eigenvectors Volume I Block 3 Interpolation Unit 9: Lagrange’s Form Unit 10: Newton's Form of the Interpotating Polynomial Unit 11: Interpolation at Equally Spaced Points Block 4 Numerical Differentiation, Integration and Solution of Differential Equations Unit 12: Numerical Differentiation Unit 13: Numerical Integration Unit 14: Numerical Solution of Ordinary Differential EquationsCOURSE INTRODUCTION Mathematical modelling of physical/biological problems generally give rise to ordinary or partial differential equations or integral equations or in terms of a set of such equations. A number of these problems can be solved exactly by using analytical methods but most of the time itis not possible to get the exact solution. Thus, a need arises to devise numerical methods to solve these problems. These methods of solving such problems may give rise to a system of algebraic equations or a non-linear equation or system of non-linear equations. The numerical solutions of these systems. of equations are quantitative in nature but when interpreted give qualitative results and are very useful. Numerical analysis deals with the development and analysis of the numerical methods. We are offering this course of numerical analysis to students entering the B.Sc. (G)/B.A. (G) as an elective course in sixth semester. Itwas in the year 1624 that the English mathematician Henry Briggs used a numerical procedure to construct his celebrated table of logarithms. The interpolation problem Was first taken up by Briggs but was solved by the 17* century mathematicians and physicists, Sir Isacc Newton and James Gregory. Later on, other problems were considered and solved by more and more efficient methods. In recent years the invention and development of electronic calculators/computers have strongly influenced the development of numerical analysis. This course assumes the knowledge of the course BMTC-131 on calculus. Number of results from linear algebra is also used in this course. These results have been stated Wherever required. For details of these results our linear algebra course (BMTE-141) may be referred. This course is divided into 2 volumes. First volume consists of 2 blocks. The first block, deals with the problem of finding approximate roots of a non- linear equation in one unknown. We have started the block with a recall of several important theorems from calculus which are referred to throughout the course. After Introducing the concept of ‘error’ that arise due to approximations, we have discussed ‘two basic approximation methods, namely, bisection and fixed point iteration methods and commonly used chord methods, namely, regula-falsi, Scant and Newton-Raphson methods. In Block 2, we have considered the problem of finding the solution of systems of linear equations. We have discussed both direct and iterative methods of solving systems of linear equations, Second volume consists of 2 blocks. Block 3 deals with the theory of interpolation. Here, we are concemed only with polynomial interpolation. The existence and uniqueness of interpolating polynomials are discussed. Several forms of interpolating polynomials like Lagrange’s and Newton's divided difference forms with error terms are discussed. This block concludes with a discussion on Newton's forward and backward difference form. in Block 4, using interpolating polynomials we have obtained numerical differentiation and integration formulae together with their error terms. After a brief introduction to difference equations the numerical solution of the first order ordinary differential equation is dealt with. More precisely, Taylor series, Euler's method of second order and Runge Kutta method of order two, three and four are derived with error terms for the solution of differential equations. Each block consists of 3 to 4 units. All the concepts given in the units are followed by a number of examples as well as exercises. These will help you get a better grasp of the techniques discussed in this course. We have used a non-programmable scientific calculator for doing computations throughout the course. While attempting the exercises given in the units, you would also need a calculator. The solutions/answers to the exercises in a unit are given at the end of the unit. We suggest that you look at ‘them only after attempting the exercises. A list of symbols and notations are also given in for your reference. Now, a few words about the way you should study the materiall We have presented this course with the assumption that you have already studied the courses,‘Calculus’ and ‘Linear Algebra’. Do not simply read the material in these blocks. You must actually Interact with every line, idea, example and question in it. As you know, Whenever we introduce concepts, we give you a lot of concrete examples to help you Understand it. We also include a list of exercises to help you strengthen your Understanding of the concepts and processes concerned. You must solve every exercise as you come to it, fo benefit from it Now, @ word about the layout of a block. In each block, you will first find a block introduction, followed by a lst of symbols that are used in the block. And then come the units of the block, Every unit starts with an introduction, where we also list the precise learning objectives of the unit. Each unit has been divided into sections. Since the material in the different units is heavily interlinked, we will be doing a lot of cross referencing. For this purpose we will be using the notation Sec.x.y to mean Section y of Unit x. As in your earlier mathematics courses, you will find the examples, ‘exercises and important equations numbered sequentially throughout a unit. Further, the exercises in each unit are interspersed within the text, They are meant to help you check your progress. The solutions, or solution outlines, or answers to the ‘exercises in a unit are given at the end of the unit, After you finish studying a unit, please go back to the objectives of the unit (given in the introduction of the unit), and see if you are confident that you have achieved them. AAs part of the tutorial component, at the end of every block you will find several miscellaneous examples and exercises for you to do. These are based on all the units you would have studied upto that point. Further, we will send you an assignment. Itis meant to be a teaching ald, apart from an assessment aid. Your academic counsellor, at the study centre, will assess this, and return it to you with suitable detailed remarks. You may like to look up some more books on the subject and try to solve some exercises given in them. This will help you get a better grasp of the techniques discussed in this course. We are giving you a list of tiles which will be available in your study centre for reference purposes. Also, we are listing a few useful websites where. you can access the content related to your course. Some Useful Books 1. Numerical Methods for Scientitic and Engineering Computation by M.K. Jain, S.R.K. lyengar, P.K. Jain, from New Age International Publishers. 2. Elementary Numerical Analysis: An Algorithmic Approach, by Samuel D. Conte and Garl De Boor, McGraw Hill Publications. Some Useful Websites 1, _hitps/freevideolectures. com/course/3597/numerical-analysis 2. https:/nptel.ac.inicourses/111/107/111107105/ 3. _hilps:/ocw.mit,edu/courses/mathematics/18-335}.introduction-to-numerical- methods-spring-2019/ 4. https:/ocw.mi,edu/courses/mathematics/18-330 introduction-to-numerical- analysis-spring-2012/ Wishing you a happy learning experience! The Course Teamlignou BMTE-144 ‘Gerciny NUMERICAL ANALYSIS Indira Gandhi National Open University ‘School of Sciences Block 1 SOLUTIONS OF NONLINEAR EQUATIONS IN ONE VARIABLE Block Introduction 7 Notations and Symbols 8 UNIT 1 Errors and Approximations 9 UNIT 2 Iterative Methods for Locating a Roots 4 UNIT 3 Chord Methods for Finding Roots 71 UNIT 4 Approximate Roots of polynomial Equations 99 Miscellaneous Examples and Exercises 17Course Design Committee Prot. Rashmi Bnardw) G.GS. Inéraprasina Unversity, Dein! Dr, Sunita Gupta LS.R. Colege, University of Dein Prot, Amber Habib ‘Shiv Nader University ‘Gautam Buddha Nagar, U.P Prot. S.A. Katro University of Pune, Pune Prot. V. Krishna Kumar NISER, Bhubaneswar Dr, Amit Kushreshha HSER, Mohal Dr. Aparna Menra LT. Delhi Prot. Rahul Roy Inaian Statistical institute, Dein Prof. Meena Sahai University of Lucknow Dr. Sachi Sivastava University of Dein Faculty Members ‘School of Sciences, IGNOU Prof. M. S. Nathawat (Director) Dr. Deepika Mr, Pawan Kumar Prof, Poorrima Mal Prof, Parvin Sinciair Prof, Sujatha Varma Dr. 8. Venkataraman ~The course design Is based on the recommendations ofthe Programme Expert Committee and the UGC: BCS template Block Preparation Team Prof. Poomima Mital (Ector) (Rota) ‘School of Sciences IGNOU, New Deli Course Coordinator: Prof, Deepika ‘School of Sciences IGNOU, New Dein Prof. Deepika and Dr. Pawan Kumar Acknowledgement: To Sh. Santosh Kumar Pal and Sh. Prashant Kumar for the word [processing and to Sh. S. S. Chauhan for preparing CRC of this block. Parts of this block are ‘based on the elective course Numerical Analysis (MTE-10) of the Bachelor's Degree Programme. Diselaimer ~ Any materials adapted from web-based resources in this course are being used for educationl purposes ‘only and not for commercial purpose November, 2021 (© Inara Gandhi Netonal Open University, 2021 ISBN-978-99-80200-90.3, Allright reserved. No part ofthis work may be reproduced in any form, by mimeograph or any other ‘means, without permission in wring trom he Indra Ganahi National Open University. FFurtner ntormation on the indra Gandhi National Open Universty courses, may be obtained from the University’s office at Matéan Garhi, New Deln-1 10 068 and IGNOU website wwyw.ignou.2c. Printed and published on behalf ofthe Indra Gandhi Netional Open University, New Delhi by Prot. Sujatna Varma, Director, School of Sciences.BLOCK INTRODUCTION This is the first of the four blocks which you will be studying in the Numerical Analysis course. In this block we shall be dealing with the problem of finding approximate roots of a non-linear equation in one unknown. In the first block of the Calculus course you have studied some methods for solving polynomial equations of degrees up to four. In this block we shall introduce you to some numerical methods for finding solutions ot equation. These methods are applicable to algebraic and transcendental equations. This block consists of four units. In Unit 1, we begin with a recall of important theorems from Calculus which are referred to throughout the course. We then introduce you to the concept of ‘error that arise due to approximation. We shall acquaint you with two types of errors that are common in numerical approximation. In Unit 2, we shall discuss two basic approximation methods, namely, bisection method and fixed point iteration method. Each of these methods involve a process that is repeated until an answer of Tequired accuracy is achieved. These methods are known as iteration methods. We shall also discuss three chord methods, namely, secant method, regula-falsi method and Newton-Raphson method in Unit 3. Unit 4, which is the last unit of this block, deals ‘with the solutions of the most well-known class of equations, the polynomial equations. For finding the approximate roots of polynomial equations we shall discuss Birge-Vieta method As already mentioned in the course introduction, we shall be using a non- programmable scientific calculator for doing computations throughout the block, While attempting the exercises given in this block, you would also need a calculator. We therefore suggest you fo use the non-programmable scientific calculator. Lastly we remind you to go through the solved examples carefully, and to attempt all exercises in each unit and all exercises given in miscellaneous examples and exercises. This will help you to gain some practice over various methods discussed in this block.NOTATIONS AND SYMBOLS € belongs to > contains e) greater than (greater than or equal to) R set of real numbers é set of complex numbers a! n(n=1)...3.2.1 (In factorial) {1 closed interval Il ‘open interval |x] absolute value of a number x ie, that is ata; tata, xa xtends toa lim f(x) limit of f(x)as xtends to a P(x) nth degree polynomial f(x) derivative of f(x) with respect to x £%) nth derivative of F(x) with respect tox 6 approximately equal to a alpha 6 beta Y gama e epsilon x pi = capital sigma t zotaUNIT 1 ERRORS AND APPROXIMATIONS | Structure Page No. 1.1 Introduction 9 Objectives 1.2 Essential Preliminaries 10 4.3. Round-off Error 21 1.4 Truncation Error 25 1.5 Summary 33 1.6 Solutions / Answers 34 1.1 INTRODUCTION Any mathematical problem can either be solved by an analytical method or a numerical method. Mostly, the analytical solutions are expressed in the closed ‘forms and are exact. However, most of the numerical methods give answers that are approximations to the desired solutions. in such situations, itis important to measure the accuracy of the approximate solution compared to ‘the actual solution. To find the accuracy we must have an idea of the possible errors that can atise in computational procedures. In this unit, we shall introduce you to different forms of errors which are common in numerical computations. You are already familiar with some fundamental theorems about continuous ‘unctions from your calculus course (BMTC-131), Here we shall review some of the theorems given in that course, namely, Intermediate value theorem, Rolle's theorem and Lagrange’s mean value theorem etc. We shall discuss some more theorems about the roots of the polynomial equation fram the course calculus. As you go along the course, you will be using these concepts, therefore we will discuss these theorems in Sec.1.2 under the heading essential preliminaries. We shall also discuss a few basic ideas and concepts regarding errors considerations, absolute and relative errors, round-off errors and truncation errors. We shall discuss round-off error and truncation error in, Sec.1.3 and Sec. 1.4, respectively. The basic ideas and results that we have illustrated in this unit will be used often throughout this course. So, we suggest you go through this unit very carefully.Block 4 Solution of Nonlinear Equations in One Variable Now we wil list the objectives of this unit. After going through the uni, please read this list again and make sure that you have achieved the objectives. Objectives Aiter studying this unit you should be able to ‘+ apply intermediate value theorem, Rolle’s theorem, Lagrange’s mean value theorem, Taylor's theorem and other theorems wherever applicable for continuous functions defined on closed intervals: ‘* define the term ‘error’ in approximation; ‘+ distinguish between round-off error and truncation error and calculate these errors as the situation demands. 1.2 ESSENTIAL PRELIMINARIES We shall begin the unit with a brief discussion on some fundamental theorems in calculus, namely, intermediate value theorem, Rolle's theorem, Lagrange's ‘mean value theorem, Taylor's theorem and some other theorems. You may recall from your knowledge of the course calculus (BMTC-131) that these theorems use properties of continuous functions defined on closed intervals [a,b] Let us discuss these theorems one by one. We begin with intermediate value theorem. According to intermediate value theorem, if function fis continuous on a closed interval [a, b]. then it takes on every intermediate value, i.., every value lying between f(a) and f(b), if f(a) # f(b), at some point within the interval, Formally, we can state the theorem as follows: Theorem 1(Intermediate Value (IV) Theorem): Let f be a function defined on a closed interval (a, b] satistying the following properties; a) _ fiscontinuous on the interval [a,b]. b) _c beanumber lying between f(a) and f(b) (Le., fa)f(b) and f(a)>e> f(b). You can also observe that the points (a, f(a)) and (b,f(b)) ie on the opposite sides of the line y=c. Further f is continuous, which implies that the graph crosses the line y=e at some point, which is (x,,c). We now illustrate examples based on IV theorem here. Example 1: Check whether the equation x° +2x?—x +3=8 has. solution in the interval [1, 2] Solution: Let f(x) = x*+2x*—x +3. We can then write the given equation as f(x) =8 and the following holds for f i) F Is continuous on |! 2] being a polynomial in x. li) £Q)=5,£(2)=17 and 8iles between F(1) and £(2) Thus, f satisfies all the conditions of Theorem 1. Therefore, there exists a number x, between | and 2 such that f(x,)=8. Thatis, the equation 8 has solution in the interval [1, 2] x4 2x? =n 43 1 Example 2: Find the value of x in 0.< x 0, in Eqn. (2). Then any point between a and a+h willbe of the form a +6h, 0<8<1 ‘Therefore, Eqn. (2) becomes faa 1 arom @) nt fa+h)=ta)+ht(@+"4'@+ Let us now make some remarks on the Taylor's theorem. Remark 3: Suppose that the function f in Theorem 4 is a polynomial of degree m. Then f°°(x)=0 for all r>m. Therefore, R,.,(x)=0 for all n>m. Thus, in this case, the mth Taylor expansion of f(x) about x, will be Bo) 655) 4 80)" fog) T m! ‘You may note that the right hand side of the above equation is simply a polynomial in (x—x,). Therefore, finding Taylor's expansion of a polynomial function f(x) about x, is the same as expressing f(x) asa polynomial in (x=x,) with coefficients from R. F(x) (+ Remark 4: You can see that the mean value theorem as a special case of Taylor's theorem. Suppose we put x, =a, x=b and n=0 in Eqn. (2). Then Eqn. (2) becomes fb) or equivalently f(b) f(a) =£() (b—a), which is the result of Lagrange’s mean value theorem. F(a) +f'(©) (b-a)Unit Errors and Approximations Let us consider an example. Example 7: Find the fourth order Taylor polynomial of f(x) =In x about Ky=l. Solution: The fourth order Taylor polynomial atti about x, =1 is given by PX) =f 46-D H+ 2 ray+ a GW pony. 2D og {usa Ec Now, f(x)=Inx fa) rayet ra x £0) £)=-6 ‘Substituting the values of (1), (1), £71), £7) and £1) in P,(x), we get (x=? 2 3 a Taylor's polynomial of f(x) =énx about x, Pua) = (x=) Which is the fourth order You may try some exercises. E5) If P, denotes the nth order Taylor polynomial then show that P(X) =F(g), PL (Gq) =F (g)s eons PO (Xp) = FC). 6) Obtain the third Taylor polynomial of f(x) =e" about x= We shall recall few more theorems from the course BMTC-131, which are elated to finding roots of the polynomial equations. You will see that we will use these theorems to locate the root in Unit 2 onwards in this block. ‘Theorem 5: Suppose that 7 =a +ib is a root of the polynomial equation P(x) =0. Then the conjugate of z, namely Z =a —ib is also a root of the equation P(x) =0, i.e. complex roots ocour in pairs. . We next give an important Theorem due to the mathematician Rene Descartes (1596-1650). ‘Theorem 6 (Descarte’s Rule of Signs): A polynomial equation P(x) =0 cannot have more positive roots than the number of changes in sign of its coetticients. Similarly P(x) =0 cannot have more negative real roots than the number of changes in sign of the coefficients of P(—x) . We denote by P(x) the polynomial obtained by replacing x by —x In P(x). . For example, let us consider the polynomial equation P(x) =0, where P(x) =x1=15x7+7x=11 19Block 4 Solution of Nonlinear Equations in One Variable x4 15x? 47x11 We count the changes in the sign of the coefficients. Going from lett to right there are changes between | and —15, between —15 and 7 and between 7 and —11. The total number of changes is 3 and hence it can have at most 3 positive roots Now we consider Pox) = (a) =15(-8)? 7-8) -11 15x? -7x-11 Here is only one change between | and —15 and hence the equation cannot have more than one negative root. We now give another theorem for locating the real roots of polynomial equation. Theorem 7: Let P(x)=0 be a polynomial equation of degree n >I. Let a and b be two real numbers with a 0 is a positive integer. Next definition gives us a measure by which we can conclude that the round- off error occurring in an approximation process is negligible or not. Definition: Let x be a real number and x" be an approximation to x. Then ‘we say that x’ is accurate to k decimal places ifBlock 4 Solution of Nonlinear Equations in One Variable 410-9 >> is used as an approximation to m= 3.14159265.... To how many decimal places the 335 enc value 775 iS accurate as an approximation to ? E10) Consider the numbers x =0.66662x10° and y=0.66649%10*, Find the difference x y using floating point representation correct upto four decimal places by (i) rounding-off and (ii) chopping. Also. compare the results, We discussed round-off error which is due to the approximate representation of a number. Sometimes there is a need to use a mathematical series/procedure upto a limited number of terms/steps, in that case, the error is due to truncating the seties/procedure. Such error is called ‘Truncation. error’. We will discuss truncation error in the following section. 1.4 TRUNCATION ERROR 1 sy Consider an infinite series ++ 44444 as "16 2 1 2 : i ‘ 1 This series is an infinite geometric series with common ratio —, which is less than 1, We get the sum of this infinite series as 1. But, if we truneate the series only to first three terms, we get +H 4 0.54025 40,125 =0.875. This produces an error of 1—0.875=0.125. Ths error is called the ‘truncation error’. Therefore, truncation error isthe difference between a truncated value and the actual value of a function ‘As you have seen above the truncation error occurs due to the truncated terms. That is, the terms which are excluded. Let us now try to estimate this, error. Consider the nth order Taylor polynomial expansion of f about x, which is £(x)=P,(x) +R,,,(X) as we discussed in Theorem 4 2sBlock 4 Solution of Nonlinear Equations in One Variable Then, R,.4(x) =f(x)=P, (x), and is called trune: Thus, Truncation error =R,,.,(X) Xo) (n+! It lim R(x) =0 for some x, then for that x (c), where 0 -M<(F(x)-P,0)) P(x) -M< f(x) 0,——\_- <1, and we get (to) 1 IRy a @l<—— ailsBlock 4 Solution of Nonlinear Equations in One Variable Now 1 canbe ge sles ete byes n nye ie., lim ntl 0. This shows that lim IR,,,,(x)I=0. ‘The above example shows that i n is suttficiently large, the value of the nth Taylor polynomial P, (x) at any point x, will be approximately equal to f(x,). the value of the given function f at X.. In fact, the remainder R,,.,(X) tells us how close the value P, (xy) is t0 f(X.). Example 12: Find the 2" order Taylor's expansion of f(x)=VI+x in ]-1. If about x =0. Also, find the bound of the error at x =0.2 Solution: Since, f(x) = vI+X . we have £(0) roo-— ro=4 Witx 2 “ogy aetthaxy al —— Ma pd ex, f= wa—Giex =F xed | xy st? me foy=zd+n) Po=> we apply Taylor's theorem for f(x) and get The error is given by R(x When x =0.2, we have R3(0.2)=—O2—_. where 0-0 we have loal+0)°" Lact ae Hence, 2y 15x10" 1R,(0.2)1s => = 0.0005 16 Hence, the bound of the error for second order Taylor's expansion about 2 is 0.0005. Example 13: Using Taylor's expansion for sin x about » approximate value of sin 10° with error less than 10”. 0, find theUnit. oo . . Errors and Approximations Solution: The nth Taylor expansion for sin x about x =0) given in Eqn. (10) is pox? Qn where x is the angle measured in radians. eps Hi (Qn+ht 08 (©), (13) Now, in radian measure, we have 10° = e radians, Therefore, by putting x = z in Eqn. (18) we get HotR, x) sint = +e al 1818 318) x} tis) | is the remainder after (n +1) terms. where R,,,{ = Us) =| cose Qn+t8 Now, R,., (z)- Itwe approximate sin 7 by Al i |, then the error introduced will be less than 10”, i.e. ia) Ps) ye =) cose cp" 2n+n!3) =|R.4{ %]|<10° lis) <107 Maximizing cose, we require that 1 (2) 197 n+!) (14) Using the calculator, we find that the value of left hand side of Ineqn. (14) for various n as given in Table 2 Table 2 a i 2 3 Left hand side of Inegn. (14) | 0.89x10° | 0.13x10" | 0.99x10" From the table we find that the ineqn. (14) is satistied for n =3. Hence, the required approximation is with error less than 1.0x107 Let us now find the approximate value of e using Taylor’s theorem. 29Block 4 Solution of Nonlinear Equations in One Variable Example 14: Using Mactaurin’s series for e* , show that e = 2.71806 with error less than 0.001. (Assume that e <3) Solution: The Maciaurin’s series for e* is xx Wa Putting x =1 in the above series, we get Lt 21° 3! Now we have to find n for which e=l+l le-P,()I=IR,,(D1-< 0.001. 1 e + We have chosen x, =0 and x= O [ £025) ie, x2 =1 or x, =41 Since x, lies in |—2, I, we don’t consider the positive value. ‘Therefore there exists only one point x,, =—I satisfying the theorem. Put x =x, in P,(xo), then we get Py(X_) =F) To calculate, P’(x,,), we differentiate both sides of Eqn. (1). Then we have 2 (Xa A= Xo) , HE) Ko 2 3 Putting x =x, on both sides of the above expression, we get PLK)= P(x) + (16) PL(x4) =F) 404 0...= F'(%Q) Note that apart from the first term, all other terms in the R.H.S. contain the factor (x-x,) and therefore when we put x = x, there terms 35 vanish.Block 4 Solution of Nonlinear Equations in One Variable By differentiating Eqn. (16) further, we get P(x.) =f(x,),1=2, r E6) The 3" Taylor polynomial of f(x) =e" about x =0 is > > Px) =FO+FO +O FTO Here f(x) Px)=e*, £7 Therefore, £(0) * and £"(x) =e =1, PO)=1, F(0)= Therefore, P(x) =1+x+4-+% and £%(0) = E7) P(x) =3x7 +x" +4x? +10x-6=0, P(-x) =—3x? =x°—4x?-10x-6=0, P(x) =0 has no negative real root, as there Is no change in the sign of the coetficients in P(—x). There is only one change in the sign of the coefficients in P(x) =0, therefore, P(x)=0 can take at most one positive real root. Here, £(0)=-6<0 fM)=12>0 Thus, the equation p(x)=0 has one positive root lying between 0 and 1, using Theorem 1. E8) P(x)=x‘+x'—2x' +4x-24=0. Here the sign of the constant term is negative where as the sign of leading coetticient is positive, Therefore, using corollary 2 of Theorem 7, P(x) =0 has two real roots, one positive and the other negative. 2 ce) 33S 113, m= 3.14159265, 3.14159292... and error 00000027 Then 410"? <0,00000027 < a 10 Therefore the approximation is accurate to 6 decimal places. E10) Here x =0.66662x10° and y = 0.66649 x 10° i) Incase of round-off upto four places 0,6666% 10° y= 0.666510" x—y=0.0001%10° 0.1% 10 il) Incase of chopping upto four places 26 x =0.6666%10°Unit. Errors and Approximations 6664%10° x—y=0.0002%10° 2x10? The exact value of x—y is 0.00013%10° close to the value in case of round-off. 0.1310", which is more E11) f@)=—-, f= Tex Po) 1) 2). (x) = DD)CM. pr) = Cyrnt (xy The function f(x) and its derivatives of different orders are continuous in ] J 1 Therefore, by Taylor's theorem Kya! Fox) yn! al i E12) No. Because the derivatives of f(x) are not defined at x = £13) 8" order Taylor expansion of f(x) =cosx about x, =0 is 37Block 4 Solution of Nonlinear Equations in One Variable Since, x ties in| -™, =|, and we have Ix1s™ <1, 4 4 therefore, we get IR, (9 1-<5 = 0.00000275573 E14) We have seen in E12) that the remainder in the 8" order Taylor expansion of cosx is such that IR, (x)! S10 ie. 1c0sx—P,(X)ISIOS [VFO =P,O) +R, 00] where P,(x)=1-4 45 -* 4% al 6! 8 This shows that we can approximate f(x) by the 8" order Taylor polynomial with error bound 10°* ie., E15) IR, n+l where c lies between () and x Therefore, IR, Ixi< 1 and le*i 5. In these cases, we need to find the approximate methods for the determination of the solution of equations of the form f(x) = 0. The problem of tinding approximate values of roots of polynomial equations of higher degree was initiated by Chinese mathematicians. The first noteworthy ‘work in this direction was done in Europe by the English mathematician Fibonacci. Later in the year 1600 scientists made significant contributions to the theory.Block 4 Solution of Nonlinear Equations in One Variable In this unit as well as in the next two units we shall discuss some numerical ‘methods which gives an approximate solution of an equation f(x) =0. We can classily the methods of solution into two types, namely (i) Direct methods and (il) iteration methods. Direct methods produce solutions in a finite number of steps whereas iteration methods give an approximate solution by repeated application of a numerical process. You wil find later that for using iteration methods we have to start with an approximate solution. Iterations methods improve this approximate solution. We will begin this unit by first discussing methods which enable us to determine an initial approximate solution in Seo. 2.2. We will discuss two iteration methods, namely bisection method and fixed point method in Sec.2.3 and Sec.2.4 respectively. We will also discuss the convergence criterion in Sec.2.5 to refine this approximate solution, Now, we will ist the objectives of this unit. After going through the unit, please read this list again and make sure that you have achieved the objectives. Objectives Ater studying this unit you should be able to ‘+ find an initial approximation of the root using tabulation method and ‘graphical method: ‘use bisection method and fixed point iteration method for finding approximate roots of the equation f(x) =0: ‘+ apply the convergence criterion to find the order of convergence of these methods. 2.2_ INITIAL APROXIMATION TO A ROOT ‘As we mentioned eartir, finding roots 6f number of equations of the form £(x)=0 is difficult. Equation f(x) =0 may be algebraic or transcendental depending on whether the function f defined by £(x) is algebraic or transcendental, respectively. Function f is called algebraic, it it has some real numbers and rational power of x. The transcendental functions includes non- algebraic functions such as a*,e*, logx, sinx, cosx, sin” x, cos x, ete. Equation 4x’—-2x+3=0 is an example of algebraic equation, whereas equations tan x —x =0, e* cosx +x* +log(2x -1) =0, etc. are transcendental. We need to find the roots of such equations. By a root of this equation we mean a number x, such that f(x,)=0. The root is also called a zero of f(x) It F(x) Is linear, itis of the form ax +b =0, a #0 and it has only one root given by Any equation which is not near is called a non-linear a equation. In this unit we shall discuss some methods for finding roots of the equation f(x) =0 where f(x) is a non-linear function. The methods which are used to find the roots of algebraic and transcendental equations of the form f(x) =Oare either iterative or direct.UM 2 cssorsnsnmnnnnnnenacrnenene eusnnnnee -nodterative Methods for Locating a Root 1. Iterative Methods (Indirect Methods): As the name suggests these ‘methods are based on the concept of successive approximation. In such ‘methods, we start with one or more initial approximation to the root and perform a sequence of iterations till the end convergence is achieved. You will see in later units that in some methods initial approximation of root lies in an interval, and in some methods initial estimate of solution is used 2. Direct Methods: These methods are used to find all the roots at the same time and give the exact value of the roots in finite number of steps. We shall discuss these methods in Block 2. In this block, we shall discuss various iterative methods to find the approximate value of the root. These methods are iterative in nature, and require some initial approximation, Thus, these methods involve two steps’ Step 4: To find an initial approximation of @ root. Step 2: To improve this approximation to get a more accurate value of the root. In this section, we will consider the Step 1 and discuss two methods — tabulation method and graphical method of finding an initial approximation toa root of an equation. Tabulation Method This method is based on the intermediate value theorem (IV Theorem), see (Theorem 1, Unit 1). We will use the theorem to locate the root in tabular form as well as graphically. Let us first try to understand the various steps involved in the tabulation method through an example. ‘Suppose we want to find a root of the equation 2x —logig x = 7 We first compute the value of f(x) =2x—log,, x~7 for different values of x say x and 4. These values are given in Table 1. Table 1 x io) 3 4 =3.301 | -1.477 | 0.397 Note that log,, x is computed using a scientific calculator. From Table 1, we find that f(3) is negative and f(4) is positive. Now we apply IV Theorem to the function f(x) = 2x —log,, x—7 in the interval I, =[3, 4]. Sinee, (3) and £(4) are of opposite signs, by IV theorem there exists a number x, lying between 3 and 4 such that f(x,)=0. Thatis, a root of the function lies in the interval 3, 4[ . Note that this root is positive Let us now repeat the above computations for some values of x lying in the interval |3, 4[ say x =3.5, 3.7 and 3.8. In Table 2 we give the values of £(x) for these values of xBlock 4 Solution of Nonlinear Equations in One Variable Table 2 |x | FOO 38 =0.544 | —0.168 | 0.0202 Here, we find that £(3, 7) and £(3, 8) are of opposite signs. By applying IV theorem again for f(x) in the interval 1, =[3.7, 3.8], we find that the root of £(x) =0 lies in the interval 3.7, 3.8[. Note that this interval is smaller than the previous interval. We call this interval a refinement of the previous interval Let us repeat the above procedure once again for the interval I. In Table 3 we give the value of f(x) for some x between 3.7 and 3.8. Table 3 x 375_[_378 3.79 Fix)_| =0.074 | =0.0175 | 0.0014 Table 3 shows that the root lies within the interval ]3.78, 3.79 and this interval Is much smaller compared to the original interval |3, 4]. The procedure is terminated by taking any value of x between 3.78 and 3.79 as an initial approximate value of the root of the equation £(x)=2x —log,, x-7=0. Let us see this example. Example 1; Find the approximate value of the real root of the equation using tabulation method Se dxT + 0. f()=14444- 7 We find that £(0) and (1) are of opposite signs. Therefore a root lies between 0 and 1. Now, to get a more refined interval, we evaluate f(x) for some values between () and |, The values are given in Table 4. Table 4 x 0 os [06 fay | -2 | 04375 | 0.4336 This table of values shows that £(0.5) and (0.6) are of opposite signs and hence the root lies between 0.5 and 0.6. We repeat the process once again for the interval [0.5, 0.6] by taking some values as given in Table 5. Table 5Unit2 erative Methods for Locating a Root From Table 5 we find that the root lies between 01.54 and 0.56. This interval is small, therefore we take any value between (1.54 and 0.56 as an initial approximation of the root. Since f(0.56) is near to zero than (0.54), we can take any number near to 0.56 as an initial approximation to the root. We will now discuss the graphical method to locate the initial approximate value of the root of an equation. Graphical Method Let us see how do we locate the root of an equation, geometrically. We draw ‘the approximate graph of y=f(x). The points where the curve cuts the x- axis are taken as the required approximate values of the roots of the equation F(xX)=0 Let us consider the equation x* —e* =0t0 locate the approximate root graphically. We first draw the graph of the function f(x) =x? -e*. You may observe that itis not easy to graph this function. if we split the function as F(x) (8) = x), where f,(x)=x", f(x) =e", then we can easily draw the graphs of the functions f(x) and f(x). The graphs are given in Fig. 1 Fig. 1: Graph of f,(x) =? and f,(x)=e* Fig.1 shows that the two curves y, =x’ and y, =e" intersect at some point P. From Fig. 1, we find that the point of intersection of the two curves is - 0.7 approximately. Thus, we have f,(-0.7) =f,(-0.7),, and therefore £(-0.7)=f,(-0.7) -f,(-0.7) =0. Hence, — 0.7 is an approximate value of the root of the equation f(x)=0. From the above example we observe the following: Suppose we want to apply the graphic method for finding an approximate root of f(x)=0. Then we may try to simply the method by spliting the equation as 3(x)=0 (1) fix) =f,(0) = where the graphs of f,(x) and f(x) are easy to draw. From Eqn. (1), we have f,(x)=f,(x). The x-coordinate of the point at which the two curves 45Block 4 Solution of Nonlinear Equations in One Variable y, =f,(x) and y, =f,(x) intersect gives an approximate value of the root of the equation f(x) =0. Note that we are interested only in the x-coordinates, we don't have to worry about the point of intersection of the curves, Often we can split the function f(x) in the form (1) in a number of ways. But we should choose that form which involves minimum calculations and the graphs of f,(x) and f,(x) are easy to draw. We illustrate this point in the following examples. Example 2: Find an approximate value of a root of the biquadratic equation xttdx' 4x? -2=0 using graphical method. Fig. 2: Graph of f(x)=x' +4x?4+4x7=2. The figure shows that the graph outs the x-axis at two points'— 2.55, and 0.55, approximately. Hence —2.55 and 0.55 are taken as the approximate roots of the equation x*+4x? + 4x? -2=0. Now go back for a moment to see Example 1. You will note that the approximate initial value of one of the roots is nearly 0.55 using tabulation method, Example 3: Find an approximate value of the positive real roots of 3x —cosx —1= (0 using graphical method. Solution: Since itis easy to plot 3x —1 and cosx, we rewrite the equation as 3x-1=cosx. The graphs of y, =£,(x)=3x—I and y, =f,(x)=cosx are given in Fig. 3. Itis clear from Fig.3 that the x-coordinate of the point of intersection of y, and Ys is approximately 0.6. Hence, x = 0.6 is an approximate value of the root of the equation 3x -cosx—1Unit2 Hherative Methods for Locating a Root ob 10 Fig. 3: Graph of f(x XL and £,(x)=cosx We now make a remark. Remark 1: You should take some care while choosing the scale for graphing. ‘A magnification of the scale may improve the accuracy of the approximate value Try the following exercises, 1) Find an initial approximation to a root of the equation 3x —/i+sinx =0 using tabulation method. 2) Find the approximate location of the root of the equation x*—4x— in the regions given using graphical method. So far, we have seen how to locate the approximate value of a root for any function. Once an initial value is decided, we use iterative methods to improve ‘the solution. In tabulation method we found that one way of improving the process is refining the internal within which a root lies. A modification of this method is known as bisection method. In the next section we will discuss this method 2.3 BISECTION METHOD ‘The method deals with finding the approximate value of the root of the equation by refining the initial interval found using tabulation method discussed earlier. ‘Once we know an interval in which a root lies, there are several procedures to refine it. The bisection method is one of the basic methods among them. AS the name suggests, in bisection method, the initial interval is bisected successively until it becomes sufficiently small. This method is also known as ”Block 1 48. Solution of Nonlinear Equations in One Variable the interval halving method. This method is not the fastest method available but tis a reliable method to find the approximate root ‘Suppose that we are given a continuous function f(x) defined on [a, b] and we want to find the roots of the equation f(x) = 0 using bisection method. We apply the following steps: Step 4 (Finding an initial bisecting interval): Find the points ap and by in the interval [a, b] such that f(a)f(bo) <0. Then by IV theorem a root lies in the interval Jag, bol - Step 2 (Finding the mid-point): Find the middle point co of the interval ag +bo Jag, bol Le, c, ~ If £(9) =0, then cp is the required root of the equation f(x) =0 and we can stop the procedure. Otherwise we go to Step 3. Step 3 (Checking the sign of the product): Find out if f(a9).f (co) <0. then the root lies in Jag,¢g{ . Otherwise the root lies in co, bo|. In either case the interval obtained is of haif the width of the original interval containing the root. We call the new interval as Ja,,b)[ Step 4: Repeat Step 2 and Step 3 with the new interval. This process either gives you the root or an interval having width 1/4 of the original interval Jaq. bo[ , which contains the required root. Step 5: Repeat this procedure unt the interval with is-as small as we desire. You may note that each bisection halves the-length of the preceding interval: Atter N steps, the original interval length will be reduced by a factor 1/2", and would be !°0- #0! 2 Geometrically, we can visualize these steps in Fig. 3. f(x) = 0 (Root) Initial interval (a) Initial interval Jay, bal containing the root of f(x) =0Unit2 Hherative Methods for Locating a Root ‘Second interval (b) The point c, and the new interval Ja,, b,[. ‘Third interval a+b, (6) The point , and the new interval Jas, bs[ Fig. 4: Bisection Method Geometricaly, itis clear from Fig.4, that we start finding root by finding aq and bby which are the end points of the interval in which a solution exists. Then, ag tb mid point cy =A is taken which gives first estimate. Further, second interval is found. Itis either Jag.cof oF Ie,.byf and is denoted as Ja,, by Again we find mid point of Ja,. b, | and repeat the process till the numerical solution is accurate enough to accept. Let us now illustrate the method throught examples. We begin with an example of logarithmic function. Example 4: Consider the equation 2x —log,, x =7. The initial root lies in the interval |3.78, 3.791 . Apply bisection method to find an approximate root of the equation correct to three decimal places. Solution: Let f(x) =2x —log,, x7. From Table 3, in Sec.2.2 we find that £(3.78) = 0.0175 and £(3.79) = 0.0014.. Thus, a toot lies in the interval 13.78, 3.79| . We apply steps of bisection method and get the following iterated values. Hteration 1: initial interval is |3.78,3.79f, therefore, ay =3.78,b, =3.79 ”Block 4 Solution of Nonlinear Equations in One Variable ag +by _GB.78+3.79) 3.785 The middle point is c, (3.785) =-0.008140 fey) = Since, f(3.78) f(3.785) > 0, the root does not lie in the interval 13.78, 3.785| . However, it lies in the interval |3.785, 3.79[ as £G.785)F(3.79) <0. So, we set a, =co =3.785 and b, =by =3.79. Iteration 2: £(¢,) =£8.7875) =-0.0034 Since, f (3.7875) f (3.79) <0, the root lies in the interval 13.7875, 3.781. Seta 3.7875and b, Iteration 3: a, =3.7875,b, =3.79 £(€9) = (3.7888) =-0.001 Since, (3.7888) £13.79) <0, the root lies in the interval 13.7888, 3.791, Now, set a3 =c) =3:7888-and by $3.79 Iteration 4: a3 =3.7888 and b, =3.79 a, +b, = 3.7894, 3.7888 43.79 =e £(€3) = (3.7894) = 0.0002 Since, f(3.7894) f (3.7888) <0, the root lies in the interval 13.7888, 3.7894. Now, set ay =a; =3.7888 and b, =c, =3.7894 Iteration 5: a, = 7888, b, =3.7894 agth, _3,7888+3,7894 2 2 (cq) =£(3.7891) =-0.00004 =3.7891 « Since, F(3.7891)f(3.7894) <0, the root lies in the interval 13.7891, 3.7894]. Root lies between ¢, and b, All these five iterations are summarised in Table 6. 50Unit2 _lherative Methods for Locating a Root Table 6 Numberof] a, | b,, | Bisected | fic, ;) | Improved Interval iteration value cy (n) 1 378 | 379 3.765 | —0.0081 13.785, 3.79] 2 376s | 378 | 3.7875 | —00034 | ]3.7875, 3.79] 3 37675 | 3:79 | 3.7686 | —0.001 13.7888, 3.79] 4 a7eee | 379 | 3.7894 (0.0002 [ 13.7888, 3.7894] 5 ‘37ees | 3.7894 | 3.7891 | —0.00004 | ]3.7891,2.7804[ ‘The width of the improved interval |3.7891, 3.7894 after 5 bisections is 3.7894 — 3.7891 = 0,003. We can say that the interval 13.7891, 3.78941 contains the root accurate upto three places of decimal. If we stop further bisections, the maximum absolute error would be (),0003. The approximate (3.7891 +3.7894) root can therefore be taken as <=") 3.7802. Hence, the desired approximate value of the root rounded-off to three decimal places is 3.789. In the following example, we will apply bisection method for trigonometric equation. Example 5: Apply bisection method to find an approximation to the positive root of the equation 2x - 3sin x 5 =0 rounded-off to three decimal places. The initial interval is ]2.8, 2.9}. and £(2.8).f(2.9)<0 Now we apply bisection method to f(x) =0 with initial interval }2.8, 2.9[ . The successive iterated values are given in Table 7. Table 7 Number] a, bay | Bisected Flea) Improved interval of value Iteration Cnt 7 28 29 2.85 0.1624 1285, 29] 2 2.85 29 2.875 = 0.0408 12.875, 2.91 3 2.875 29 2.8875 0.02089 12.875, 2.88751 4 2.875 | 2.8875 | 288125 | —0.0097 12.8125, 2.88751 5 | 288125 | 28875 | 2.884975 | 0.0056 | 12.88125, 2.843751 6 | 2.88125 | 2.884375 | 2.8828125| —0.0021 _|_|2.8828125, 28843751 7__ | 2.882125 | 2.884375 | 2.8835936| 0.0017 _| 12.8828125, 2.8835938 8 | 2.8828125 | 28835088 | 2.8832031 | 0.0002 _| 12.8832031, 2.8835938| stBlock 4 52 Solution of Nonlinear Equations in One Variable After 8 iterations the width of the interval is 2.835938 — 2.8832031 = 0,0003907. Hence, the maximum possible absolute error to the root is 0,0003907. Therefore, the required approximation to the root is 2.883 correct to three decimal places. Now, let us apply bisection method for polynomial equation. Example 6: Find a root of the equation x° —4x -8.94 decimal piaces using bisection method, 0 accurate to three Solution: Here, f(x) =x'—4x-8.95=0. First, we will find the initial interval £()=()' 4) -8.95==11.95<0 £(2)=(2)' -4(2)-8.95=-8.95<0 £(3)=(3)'-4(3)-8.95= 6.05 >0 Since f(2) f(3)<0, therefore, f(x)=0 has a root in 2,31 We compute successive iterations using |2, 3| as initial interval and show the improved intervals in Table 8. Table 8 D & Fie, _ | Mnproved interval a co 1 3 25 | -3325 2) 25 3 275 | 084688 a| 25 273 9 2.605 | — 136208 | 225, 2751 4/2625 | 2.75 | 26875 | 028911 | _12.6873, 2.751 5 | 26875 | 2.75 [2.71875 | 0.27092 | 2.6875, 2.718751 6 7 8 9 26875_| 2.7187S)| 2.70312 | 0.01108 | 1270312, L7187ST [2.70312 | 2.71875 | 2.71094 | 0.12942" | 2.70312, 2.710941 70312 | 2.71094 | 2.70708 2.70312, 2.707031 [2.70312 | 2.70703 | 2.70508 | 0.02396 | 2.70312, 2.70508] TO] 270512 | 2.70508 | 2.70410 | 0.00643 | P.70312, 2.70410, Hence, the root is 2.704 accurate to three decimal places. Let us consider the transcendental equation with exponential function. Example 7: Perform three iterations to find the root of the equation e* —3x=0 using bisection method with initial interval 1.5.1.6 Solution: Let f(x) =e" —3x=0 F(1.5) =e! -3(1.5)=-0.01831<0 £(1.6) =e! ~3(1.6) =0.15303>0 iteration 1: ay =1.5.b =1.6.Unit2 Hherative Methods for Locating a Root £(¢,) =£0.525) Since, f(1.525)-f(1.5) <0, the root lies in the interval ]I.5, 1.525 Now set a) 525 eration 3: ay =1.5,b) =1.5125 13541.525 & F(c,) =F(1.5125) = 0.00056 Since, f(1.5).£(1.5125) <0, the root lies in the interval 115, 1.5125, Now a3 =a) =1.5 and bs =¢) =1.5125 Thus the required approximation to root after 3 iterations is 1.5125, Now let us make some remarks Remark 2: While applying bisection methad we must be careful to check that f(x) Is continuous. For example, we may come across functions like f(x) 4 we consider the interval 10.5, 1.51, then £(0.5) (5) <0 In this case we may be tempted to use bisection method. But we cannot use the method here because f(x) is not defined at the middle point x =1. We can overcome these difficulties by taking f(x) to be continuous throughout the initial bisecting interval. (Note that i f(x) is continuous, by IV theorem f(x) assumes all values between the interval.) ‘Therefore, you should always examine the continuity of the function in the initial interval before attempting the bisection method. Remark 3: It may happen that a function has more than one root in an interval. The bisection method helps us in determining one root only within one interval. We can determine the other roots by properly choosing the other initial intervals. You can try some exercises now. 3) Starting with the interval [a,, ba], apply bisection method to the following equations and find an interval of width 0.05 that contains a solution of the equations a) =x =0, [ay, by] =[1.0, 1.8]Block 4 54 Solution of Nonlinear Equations in One Variable b) Inx=54x=0, fay, by]=13.2, 4.0] E4) Using bisection method find an approximate root of the equation 2x*—x—4=0 in the interval |1, 2[ correct to two places of decimal. ‘You may recall that in the introduction we classified an indirect method as an Iteration method. Bisection method is one of the iteration methods. Let us consider another iteration method now called fixed point iteration method in the following section. 2.4 FIXED POINT ITERATION METHOD The bisection method we have described earlier depends on our ability to find an interval in which the root lies. The task of finding such intervals is difficult in certain situations. In such cases we try an alternate method called Fixed Point Iteration Method. We shall discuss the advantage of this method later. The first step in this method is to rewrite the equation f(x) = x=20) @) For example, consider the equation x? — We can write it as Vixe8 ) 2 x= a8 4) x 6) We can choose the form (2) in several ways,-Sinee, f(x) =0 is the same as = 2(%) finding a root of F(x) =0 Is the same as tinding a root of x= 2(x) i.e., a fixed point of g(x). Each such g(x) given if (3), (4) or (5) is called an Iteration function for solving f(x) =0 Here is the formal definition of fixed point. Defini 2(@) In: A point cris called a fixed point of an iteration function g(x), i a. For example, consider (x)= x", you may note that g(0)=Oand g(1 Therefore, 0 and 1 are two fixed points of g(x) =x*. Note that not all equations have fixed points. For example, g(x)=x+1 has no fixed points, since x is never equal to x +1 for any real number. (Once an iteration function is chosen, our next step is to take an initial point Xp, close to the expected root, as the initial approximation of the root Starting with suitable x,,, we find the frst approximation x, as X, =B(%o) Then, we find the next approximation as Similarly, we find the successive approximations as.Unit2 Hherative Methods for Locating a Root x) =8(%2) Ky =B(X5) Kyat = 8%) Each computation of the type X,,, =2(X,,) 18 called an iteration. Now, two questions arise (When do we stop these iterations? (il) Does this procedure always give the required solution? To ensure this we make the following assumptions on g(x). While selecting (x) in an interval, say 1, containing the initial approximation. () The derivative 2’(x) of g(x) exists in I. (i) 2°69) is continuous in 1. |.a/(x) <1 in an interval containing the initial approximation to the root. Xp. (That would mean that we require | (x,,)| 1.. For any positive real number x, wwe see that the inequality (2x +8)" >1 is satisfied. Therefore, we consider any interval on the positive side of x-axis. Since the starting point is x9 = ‘we may consider the interval I =[3, 6). This contains the point 5. Now s