You are on page 1of 1

Assignment-1 Academic Year: 2019-20

Matrix Computation (MC404)

1. Consider the following floating-point number system.


Base 𝛽 = 2, length of mantissa 𝑚 = 3 and range of exponent {−1,0,1,2}. Find all the decimal
numbers which can be represented by this system.
0.1𝑎 0.1𝑎
2. Explain the ill-conditioning of a matrix. Let 𝐴 = [ ] . Determine a such that condition
1 1.5
number of 𝐴 is minimized. Use the maximum norm.
3. Consider the following system of equations.
0.0030𝑥1 + 59.14𝑥2 = 59.17
5.291𝑥1 − 6.130𝑥2 = 46.78
Solve the system using Gaussian elimination and use four-digit arithmetic with rounding. Is it a good
approximation to the exact solution? If no, then why is it so and demonstrate how can you get a good
approximation?
4. Find the condition number of the following system.
2.1 1.8 𝑥1 2.1
[ ][ ] = [ ]
6.2 5.3 𝑥2 6.2
Discuss the sensitivity of the solution of the system to roundoff errors.
1 1
5. Find orthonormal vectors 𝑞1 , 𝑞2 , 𝑞3 such that 𝑞1 , 𝑞2 span the column space of 𝐴 = [ 2 −1].
−2 4

6. Is it possible to decompose a matrix in lower and upper triangular matrices? If yes, prove it. Else give
a counterexample.
2 1 −1
7. Is it possible to find Cholesky decomposition of the matrix 𝐴 = [ 1 1 0 ] ?
−1 0 −1
2 1 −1
8. Is matrix 𝐴 = [ 1 1 0 ] positive-definite? Justify your answer.
−1 0 −1
1
1 2
9. Check if the matrix 𝐴 = [ 1 ] is positive definite. Obtain its Cholesky decomposition.
2
2
1 −1
10. Obtain QR factorization for the matrix 𝐴 = [ ].
−1 1
1 −1 1
11. Solve 𝐴𝑥 = 𝑏 for 𝐴 = [ ] and 𝑏 = [ ]. Does the solution exist? How many least-square
−1 1 0
solutions are possible? Compute least norm solution if there are many.
1 3
12. Solve [ ] 𝑋 = [ ]. Does the solution exist? If not, find the least square solution.
2 4
1
13. Compute singular value decomposition of the matrix 𝐴 = [ ].
2
1 0 1 0
14. Find the SDV of the matrix 𝐴 = [ ]. .
0 1 0 1
15. Calculate condition number of following matrices with respect to 1-norm, 2-norm, ∞-norm.
1 0 0 1
(i) 𝐴=[ ] (ii) 𝐴=[ ]
0 100 0 0
1 100 0 0
(iii) 𝐴=[ ] (iv) 𝐴=[ ].
100 1 0 0

You might also like