Vous êtes sur la page 1sur 1

IIIT-Delhi

MTH-101 Monsoon 2010


Homework 2 Linear algebra and Advanced calculus

1. A is an m × n matrix of rank r. Suppose there exists some values of b for which Ax = b has
no solution.

(a) What are all the relations that must be true between m, n and r ?
(b) Explain why the system of equations AT y = 0 has solutions other than y = 0.

2. Suppose Q is a matrix such that QT = Q−1 . Show that the columns of Q (let us call them
q1 , q2 , . . .) are orthogonal unit vectors. (That is if you take any two distinct columns and take
their inner product then it will be zero. Further, the innner product of a column with itself
will be 1.)

3. If V and W are subspaces of R5 and dim(V ) = dim(W ) = 3 then prove that V ∩ W contains
a non-zero vector.

4. If Ax = b has infinitely many solutions, why is it impossible for Ax = b0 to have only one
solution (where b 6= b0 ) ? Could Ax = b0 have no solution ?

Suggested programming based problems


 
2 1
1. Make the matrix A = in Matlab by the command:
6 3
>> A = [2 1; 6 3]
Then compute b = Ax for 100 random x vectors by the command:
>> br = A * rand(2, 100);
Plot these b vectors as black dots by the command:
>> plot(br(1,:), br(2,:), k.)
What is the pattern, and why?

2. Use the Matlab command A = rand(10,5); to make a random 10 × 5 matrix A, and B =


rand(5,9); to make a random 5×9 matrix B. Then use the command [R,p] = rref(A*B);
to find the reduced row echelon form R and a list p of the pivot columns. Using this infor-
mation, give bases for the nullspace, column space, and row space of AB.

3. In Matlab, make a random 5 × 5 symmetric matrix A and copy it to B by the commands:


A = rand(5,5); A = A * A; B = A;
Now, repeatedly compute the QR factorization B = QR and then replace B with the new
matrix RQ, via the commands:
[Q,R] = qr(B); B = R * Q;
Repeat the above line over and over (you can use the up-arrow key in Matlab to fetch the
previous command), until B stops changing. You can ignore tiny numbers smaller than 10−7
(which Matlab prints as 1e-7) or so. You should find that B converges towards a diagonal
matrix ! Compare the numbers on its diagonal [diag(B) in Matlab] to the eigenvalues of A
[computed by eig(A) in Matlab].

Vous aimerez peut-être aussi