re 1 2022 MATHEMATICS — HONOURS Paper : CC-14 (Numerical Methods) Full Marks : 50 The figures in the margin indicate full marks. Candidates are required 10 give their answers in their own words as far as practicable. Each of the following questions has four possible answers of which exactly one is correct. Choose the correct alternative 1x10 (a) The relative percentage error in approximating x by 3.142 instead of 3.14159 is @ 0.001 i) 0.00041 (ai) 0.000131 Gv) 0.013 (©) Suppose a digital computer uses a number base 10, 4-bit mantissa, 2-bit exponent and 2 bits for sign, ‘one for mantissa and one for exponent, in normalized form for floating numbers. Then which one of the following is a machine number of that computer? @) 3.143110 (ii) 10400 i) 10001 Gv) 3.219710? (©) Consider the interpolating polynomial for f(x)=x3 based on the points (1, 1), (2,8), (3,27) and (4, 64). Find an upper bound of the interpolation error on 1 Sx <4. a 13 On ® 4 a = (0 (@ The trapezoidal rule of integration, when applied to | f(x)d will give the exact value of the integral @ if f(x) is a quadratic function of x Gj) if f(x) is a linear function of x (ii) if, f(x) is a cubic function of x (iv) if f(x) is any function of x (e) If an iterative method approximately cubes the error in every three iterations then the order of convergence of the method is 1 O53 Gi) V3 Gi) 3 (iv) 3 Please Turn OverQ) (xe Sm,)-Mathematics-H/CC- js true? ax+3 et Tet lh gyt 5x + 10 oft and fin) None of these. what is the best choice of second point fy (Which of the following statements (0) Simpson's rule is exact for fe) (Qt) Simpson’s rule is exact for fls)= (am Simpson's rule is exact for {@)= @L and Il ii) 1 (@) Ifit is provided that 3) = 4 is on© of the initial points, solving by Bisection Method? ( -S such that f(-5) = -%6 Gi) -3 such that f(-3) => (h) Ifyou have to solve Ax=b many times for different G@ compute the inverse of A Gi) use Gaussian elimination with partial pivoting fi) use an iterative method Gv) use LU decomposition method. Gi) 0 such that ‘f= 5 civ) 13 such that (13) =2- ‘veotors b but the same matrix Ay itis best ty _y. (0) = Applying second order Runge-Kutta method (@ Consider the initial value problem * with step length f the value of y(H) will be @ 220" -2h-2) Gi) yo - w 20(H?-2h=2) ws) 7 -1y () Let f:(0,2] > R bea continuously differentiable function with js (x)de -2f0). 0 ‘Then the error of the approximation is @ fo for some § € (0,2) i) fa for some & €(0,2) (i) FE torsomese(02) wy) £e fu tonete(0.2)- Unit ~ 1 and Unit — 2 (Answer any one question) 2. Us ’s Interpolating polynomial, find the miss le ising Lagrange’s Interpolating polynomial, fi issing term in the foll bl following table : = 0 i 5 fe) | 41 B 7 7 4 2 30 | 60 3. What are the basic features of Hermite ii res of ite interpolation formula? Prove the uni fit uniqueness of it1. 10. @) Unit - 3 (Answer any two questions) Integrate the Lagrange’s interpolation polynomial x4 xx (x)= —* flap) + F(a) xo-% 5% over the interval [x, x] and establish the trapezoidal rule. Give a comparison between Simpson’s 1/3 rule and Simpson’s 3/8 rule. HO What do you mean by degree of precision of a mechanical quadrature formula? Prove that a necessary and sufficient condition for an (+ 1) point quadrature formula with node set {%9, x1, X3y--%n} to have degree of precision (2n +1) is that 2 Jr@o,yar= + Where w(x) =(x—x9)(¢—x,).¢—4,) and Q,(x) is any arbitrary polynomial of. degree 0) is given by k-1 2 q Sn that ,41= 2 , Zest Tava] and also deduce that Gyyi=-T7r=ey, where «,, isthe error atthe mth iteration, What is the order of this iterative method? “a Explain whether one ean use Newton-Raphson method to find real root i (x)= 0 has a multiple real root? How it can be generalized and what will be its order of convergence? DOH Find the root of the equation xe*= cosx using Regula-falsi method correct to three decimal places. 5 Please Turn Over“ a simple real root cof an equation of 11. (a) Explain the method of fixed point iteration for approxi: an interval about 0 the form x=6(x), where 4(2) and 4 (0) are continuous on | of the above method. (a) #0. 24 (©) Find also the order of convergence of the above method if oo 24 | (b) Derive a sufficient condition of convergence Unit - 5 (Answer any two questions) 1(1)m let x be the Ath iterated value ofthe variable | 12. Fora system of n equations in n unknown x; (? x,a(0 =(s(®, 200,14) ne the solaton veto obtained atthe Ah iteration and § = (away be the exact solution of the system of equation. If a,,#0 for i= (Im and then, prove that the estimate of the absolute error of Gauss-Seidal iteration method can be represented | by WE-x) I< a 11¢-x4 |, where|.l| denotes a norm of a vector and aj, be the coefficient of | | i the equation in j variable x, 5 ray 13. It is given that 2, =6 is the largest eigenvalue and X, Jo 5) is the corresponding eigenvector of 0. [-306 -198 426] =| 104 67-147 | [+176 -114 244 | Find the next dominant eigen-pair of A. | ol | 14. Show that the trix E = fee eee { 1 ; do not have LU decomposition. Suppose LU decomposition of the | matrix A= (4), €Xists, then determine the 2 equations in n? eee ron ee Ce Aled fe Op, Yoreahy)4 then discuss how to solve : 2H i‘3) 15. Write down the basic assumptions for fi inding the dominant i f Power method. How the convergence rate er eigenvalue of a real matrix Ayo by eigenvalues? State when the method fail,“ #04 depends upon the magnitudes a Unit - 6 (Answer any one question) 16. Deduce Euler's method to solve inital value problem y'=/(x, ), yOxq) = using Taylor's series expansion. Does this method always converge? Justify your answer. #2 17. Solve the equatio a a (0) ak +¥°, 20) by fourth order Runge-Kutta method, from x=0 to x= 0.2, with step length h= 0.