You are on page 1of 5

Practice Exercise 1

Numerical Analysis (MT-243)

Attention: Work out all the details of your work in all of the following exercises. Give
complete reasoning wherever required.

References:
(1) Advanced Engineering Mathematics by Erwin Kreyszig(10th Edition)
(2) Numerical Analysis by Burden and Faires(9th Edition)
(3) Numerical Methods for Engineers by Steven C. Chapra(7th edition)

1 Review of Numerical Analysis


I. Difference between Numerical and Analytical Techniques
II. Advantages and disadvantages of Numerical Methods

2 Closed domain methods


 Solve the following non-linear equations by bracketing the roots(upto 4 d.p). Use
both Methods (Bisection+False position) and compare the results.
i. 𝑥3 − 𝑥 − 1 = 0 Ans 1.3125
ii. 4
4𝑥 = 𝑥 + 8 Ans 1.2330
iii. 3𝑥 − 𝑐𝑜𝑠𝑥 − 1 = 0 Ans 0.60715
iv. 𝑥
𝑒 − 3𝑥 = 0 Ans 1.5625
v. 𝑥𝑙𝑜𝑔10 𝑥 = 1.2 Ans 0.4875
 For the problem 𝑓(𝑥) = 2 − 𝑥 + 𝑙𝑛𝑥 = 0, estimate how many iterations of Bisection
method are required if we start with x0=2 and x1=4 and desire an accuracy of
𝐸 tol= | 𝑥0 − 𝑥1| ≤ 0.001.

Note: Where 𝐸 tol means how much error you can tolerate (tol).

3 Open domain methods


 Find the roots of the following nonlinear equations by Newton Raphson Method.
i. 𝑥 2 − 5𝑥 + 2 = 0 𝑢𝑝𝑡𝑜 5 𝑑. 𝑝 𝑛𝑒𝑎𝑟 𝑥0 = 0 Ans 0.43845
ii. 𝐶𝑎𝑐𝑢𝑙𝑎𝑡𝑒 𝑡ℎ𝑒 𝑟𝑜𝑜𝑡 𝑜𝑓 𝑡ℎ𝑒 𝑎𝑏𝑜𝑣𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝑏𝑦 𝑡𝑎𝑘𝑖𝑛𝑔 𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑔𝑢𝑒𝑠𝑠 𝑥0 = 4.
iii. 𝑥 3 − 2𝑥 + 2 = 0 𝑢𝑝𝑡𝑜 3 𝑑. 𝑝 𝑛𝑒𝑎𝑟 𝑥0 = 0.2 Ans -1.769
iv. 𝑥 2 + 4𝑠𝑖𝑛𝑥 = 0 Ans -1.9338
v. 𝑥 = 2𝑐𝑜𝑠𝑥 (𝑡ℎ𝑒 𝑟𝑜𝑜𝑡 𝑙𝑖𝑒𝑠 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 1 𝑎𝑛𝑑 2)
2
 The function 𝑓(𝑥) = 4𝑥 3 − 1 − 𝑒 𝑥 has values of zero near x=1 and x=3.
a) What is the derivative of f?
b) If you begin Newton’s method at x=2, which root is reached? How many iterations
to achieve an error less than 10-5?
c) Begin Newton Method at another starting point to get theother zero.
d) For both parts (b) and (c), Tabulate the number of correct digits at each iteration.

 Repeat the process of all equations given in Newton’s Section foe Secant
Method and compare the results.

4 System of Linear Equations (Direct Methods)


1. Given system of linear equations:

2. Use Gauss Jordan Elimination method and Gauss Elimination Method to solve
the system of linear equations.

i. 𝑥1 + 3𝑥2 + 4𝑥3 = 3
2𝑥1 + 7𝑥2 + 3𝑥3 = −7
2𝑥1 + 8𝑥2 + 6𝑥3 = −4 Ans (4, -3, 2)

ii. 2𝑥1 + 8𝑥2 − 4𝑥3 = 0


2𝑥1 + 11𝑥2 + 5𝑥3 = 9
4𝑥1 + 18𝑥2 + 3𝑥3 = 11 Ans (2, 0, 1)

iii. 2𝑥2 + 6𝑥3 = 2


3𝑥1 + 9𝑥2 + 4𝑥3 = 7
𝑥1 + 3𝑥2 + 5𝑥3 = 6 Ans (7, -2, 1)

iv. 𝑥1 + 3𝑥2 + 2𝑥3 + 5𝑥4 = 11


−𝑥1 + 2𝑥2 − 2𝑥3 + 5𝑥4 = −6
2𝑥1 + 6𝑥2 + 4𝑥3 + 7𝑥4 = 19
5𝑥2 + 2𝑥3 + 6𝑥4 = 5 Ans (5, -1, 2, 1)

 Use LU-Decomposition Methods i.e Crout’s factorization and Doolittle’s


Method to solve the system and compare the results.

i. 9𝑥1 + 3𝑥2 + 3𝑥3 + 3𝑥4 = 24


3𝑥1 + 10𝑥2 − 2𝑥3 − 2𝑥4 = 17
3𝑥1 − 2𝑥2 + 18𝑥3 + 10𝑥4 = 45
3𝑥1 − 2𝑥2 + 10𝑥3 + 10𝑥4 = 29 Ans (1, 2, 2, 1)

ii. Cosider the linear system

(a). Solve the linear system AX = B by using the Doolittle method.


(b). Solve the linear system AX = B by using the Crout method.
(c). Solve the linear system AX = B by using the Cholesky method.

Compare the results of all three methods, your answer must be same for all three methods.
Also verify your answer by substituting back to given equations.
Answer:

NOTE:
The only requirement to apply Crout’s Method and Doolittle’s Method is that your matrix of
co-effiecient must be square matrix. But to apply the Cholesky’s method you matrix of
coefficient must be symmetric i.e A=AT and matrix A must be positive definite. If these two
conditions hold for A, then you can apply Cholsky’s Method.
5 System of Linear Equations (Numerical Methods)
Use both the iterative techniques (Gauss Jacobi method + Gauss Seidal method) to solve the
the given system of linear equations upto ten iterations.

i. 10𝑥 + 𝑦 − 3𝑧 = 8
𝑥 + 9𝑦 − 4𝑧 = 6
𝑥 + 𝑦 + 5𝑧 = 7

ii.
Instructions

Do complete literature survey of these topics from any


reference book to completely comprehend the topics and
pick up practise work from those books too.

You might also like