3251 hw3 Completed

You might also like

You are on page 1of 4

Homework 3

Claire Liu cl3944


February 20, 2020

Question 1
Part a
If the augmented matrix is n x (n+1), then its complexity for Gaussian elimination is O(n3 ) since there is
one outer loop that iterates over the n loops ( for i in range(0,m) ), a nested loop within that loop that
performs a loop on each column of each of the n rows ( for j in range(i+1,m) ), and the statement within the
loop is a row swapping operation which costs n ( A[[i,j]] = A[[j,i]] ). When you multiply all these iterations
of n together, you get an n3 complexity. For the subsequent lines (26 - 35), the cost would also conclude
in a n3 complexity since there is the same n iterations through the loops, another n iterations through the
columns, and n operation cost for row operations in row reducing. Thus, the overall complexity is O(n3 ).

Part b
The complexity to reduce down to row echelon form is O(n). This is because there is no need for swapping
rows so we just need to iterate through the rows and reduce the one below the current pivot row; this means
that as the size of the matrix grows infinitely, the row reduction process becomes constant.

1
Question 2
Part a
α 6= 0, β ∈ R: as long as α is not equal to 0 and β is any real number, the system will have a unique solution.

Writing out the equations yields:


x1 = x3
αx2 = β
3x3 = 1

Solving these equations yields the solution of:


β
x1 = 31 , x3 = 13 , x2 = α

Part b
α = 0, β ∈ R, β 6= 0: if α is equal to 0 and β is any real number that’s not zero, the system will have no
solutions. The inconsistent equation that arises is 0x1 + 0x2 + 0x3 = β

Part c
α = 0, β = 0: if α is equal to 0 and β is equal to 0, then the equation will have many solutions. The free
variable is x2 and the solutions of the pivot do not depend on it.

Part d
-1
To
 find the first column of A , we first use Gaussian elimination on A so that it becomes:
1 0 −1
0 α 0
0 0 3
To find this matrix, I flipped the first row with the third row. Then, I matrix multiplied this matrix
with an arbitrary u1 vector which should yield e1 . By doing this, I found that the first element of u1 - the
third element of u1 would yield
 1; α * the second element of u1 = 0; 3 * the third element of u1 = 0. This
1
means that the vector u1 = 0.
0

2
Question 3
Part a
Rot2 (θ) is not a subspace of M2x2 because it doesn’t contain the zero matrix. There is no value of theta that
would lend that sin and cos both return zero.

Part b
The symbol corresponds to matrix multiplication. The * symbol corresponds to matrix exponentiation.

Part c
 
1 0
For the zero matrix: 0 * Rot2 (θ) = Rot2 (0 ∗ θ) =
0 1  
cos(−θ) −sin(−θ)
For the additive inverse matrix: -1 * Rot2 (θ) = Rot2 (−1 ∗ θ) =
sin(−θ) cos(−θ)

Part d
The axioms that are violated are the communicative axiom [u+v = v+u] and the distributive axiom [c(u+v)
= cu + cv].

Part e
The subspace spanned by Rot2 (θ) is all rotation matrices.

3
Question 4
Part a
The kernel of this transformation is the set of vectors where all the elements are the same value. This
is because all of these vectors would become the zero vector after the transformation since the difference
between each element would be equal to the average of the rest. The kernel is a subspace of Rn , any n vector.

Part b
   
2 1
If x is equal to then the output y is equal to . This indicates that the vector, y, is in the range of
1 −1
f. I got this answer using the transformation formula.

Part c
 
1
The basis for the null space of A is the 1’s vector 1 because inputting that vector would yield the zero
1
vector as our solution and it would span every vector in the null space. It is linearly independent because
there is only one vector.
 
α
To answer the second part of part c, we can say if α is an arbitrary linear combination of our basis
α
−1 −1 α α
      
1 2 2 α α− 2 − 2 0
vector, then  −1
2 1 2 α =  −α
−1   
2 +α− 2
α
= 0
−1 −1 −α α
2 2 1 α 2 − 2 +α 0

Part d
−1 −1 −1 −1 −1 −1
     
1 2 2 1 2 2 1 2 2
Gaussian elimination of A:  −1
2 1 −1 
2 → 0 3
2
−3 
2 → 0 3
2
−3 
2
−1 −1 −1 −1
2 2 1 2 2 1 0 0 0
The pivot columns indicate how to find thebasis of Col(A) so since the pivot columns are the first and
−1

1 2
second columns, the basis for Col(A) is  −1
2
 and  1 
−1 −1
2 2

You might also like