Difficult Matrix Identity Question

In summary, when evaluating the statements for all n x n matrices A, B, C, we find that (a) is not always true, as shown by a specific choice of A and B. (b) is also not always true, as it relies on the commutativity of matrix multiplication, which is not always guaranteed. (c) is false, as it assumes that A and B are scalars and not matrices. And finally, (d) is also false, as it relies on the false assumption that matrix multiplication is commutative, which is not always the case. Therefore, we can conclude that none of the statements are universally true for all n x n matrices A, B, C.
  • #1
negation
818
0

Homework Statement



For each of the following statements, select whether the statement is true or false for all n × n matrices A, B , C. (Note that you are being asked whether the statement is true or false for all n × n matrices A, B, C, not just for some A, B, C.)

a) (-6 A - 4 B)2 = 48 AB + 36 A2 + 16 B2


b) (6 I + A2)(-4 I + A2) = (-4 I + A2)(6 I + A2)

c) A = ±8 B ⇒ A2 = 64 B2

d) (A(-4 I - 6 A))T = AT(-4 I - 6 A)T

The Attempt at a Solution



What I know is (b) and (d) are false.

(b) is false because AB is not equal to BA unless the matrix is invertible.

(d) is false because (AB)T = BTAT.
 
Physics news on Phys.org
  • #2
In a and b, it's a good idea to use the distributive rules ( A(B+C)=AB+AC and (A+B)C=AC+BC ) to evaluate both the left-hand side and the right-hand side. Have you tried this?

Good thinking in d, but this is only a reason to suspect that d is false. It doesn't prove it conclusively. What do you get if you evaluate the right-hand side using the distributive rule and that ##(A+B)^T=A^T+B^T##? Is the answer different if you first change the order of the two factors on the right-hand side?
 
  • #3
Fredrik said:
In a and b, it's a good idea to use the distributive rules ( A(B+C)=AB+AC and (A+B)C=AC+BC ) to evaluate both the left-hand side and the right-hand side. Have you tried this?

Good thinking in d, but this is only a reason to suspect that d is false. It doesn't prove it conclusively. What do you get if you evaluate the right-hand side using the distributive rule and that ##(A+B)^T=A^T+B^T##? Is the answer different if you first change the order of the two factors on the right-hand side?

No I havent. How should I go about evaluating (a) and (b) using the principle you've provided.

(a): (-6 A - 4 B)2 = 48 AB + 36 A2 + 16 B2

(-6 A - 4 B)(-6A -4B) =? 48AB + 36AA + 16BB

How should I take it further?
 
  • #4
negation said:
No I havent. How should I go about evaluating (a) and (b) using the principle you've provided.

(a): (-6 A - 4 B)2 = 48 AB + 36 A2 + 16 B2

(-6 A - 4 B)(-6A -4B) =? 48AB + 36AA + 16BB

How should I take it further?
You don't. This is where you stop, because what you just did proves that a is true. (Edit: My bad. See the correction below). You could however have included more information (all the steps) in your calculation:
$$(-6A-4B)^2= (-6A-4B)(-6A-4B) = (-6A)(-6A-4B)+(-4B)(-6A-4B)=\cdots$$ You should do problem b this way. You don't have to tell me all the steps, but you should evaluate both the left-hand side and the right-hand side this way, and post the results. (If the result is wrong, I may ask you to post the calculation).
 
Last edited:
  • Like
Likes 1 person
  • #5
Fredrik said:
You don't. This is where you stop, because what you just did proves that a is true. You could however have included more information (all the steps) in your calculation:
$$(-6A-4B)^2= (-6A-4B)(-6A-4B) = (-6A)(-6A-4B)+(-4B)(-6A-4B)=\cdots$$ You should do problem b this way. You don't have to tell me all the steps, but you should evaluate both the left-hand side and the right-hand side this way, and post the results. (If the result is wrong, I may ask you to post the calculation).

Alright. Awesome. So (a) is true. You said I could have included further information (a more rigourous manner), I wish to know how.

(b): (6 I + A2)(-4 I + A2) = (-4 I + A2)(6 I + A2)

(6 I + AA)(-4I + AA) = (-4I + AA)(6I + AA)

but..

(24II + 6AAI -4IAA + AAAA) = 24I + 6AA - 4AA + A4


AB =/= BA

so (b) is false


Edit: I just saw.

(-6A-4B)^2= (-6A-4B)(-6A-4B) = (-6A)(-6A-4B)+(-4B)(-6A-4B) = 36AA + 24AB + 24AB +16BB = 36AA + 48AB +16BB
 
Last edited:
  • #6
Fredrik said:
You don't. This is where you stop, because what you just did proves that a is true. You could however have included more information (all the steps) in your calculation:
$$(-6A-4B)^2= (-6A-4B)(-6A-4B) = (-6A)(-6A-4B)+(-4B)(-6A-4B)=\cdots$$ You should do problem b this way. You don't have to tell me all the steps, but you should evaluate both the left-hand side and the right-hand side this way, and post the results. (If the result is wrong, I may ask you to post the calculation).

I think it's good.
 
  • #7
negation said:
So (a) is true.
Oops. It's not. I didn't think it through. Sorry about that. 24AB+24BA (which appears in the left-hand side when you rewrite it as a sum of four terms) is not always equal to 48AB (which appears in the right-hand side). So we can conclude that (a) is very likely not true for all A,B. To complete the proof, you should look for a specific choice of A and B that makes the equality in (a) false.

negation said:
You said I could have included further information (a more rigourous manner), I wish to know how.
Your book should include a definition of matrix multiplication, a definition of the transpose operation, and several theorems that state elementary results like ##(A+B)C=AC+BC## and ##(A^T)^T=A##. One way to include more information is to use only one of those theorems at a time. But if you want to do things like (A+B)(C-D)=AC-AD+BC-BD in one step, that's OK too. You just need to make sure that your readers can tell what you're doing.

negation said:
(b): (6 I + A2)(-4 I + A2) = (-4 I + A2)(6 I + A2)

(6 I + AA)(-4I + AA) = (-4I + AA)(6I + AA)

but..

(24II + 6AAI -4IAA + AAAA) = 24I + 6AA - 4AA + A4AB =/= BA

so (b) is false
I'm not sure sure what you're doing here. If you just take the left-hand side ##(6I+A^2)(-4I+A^2)## and rewrite it as a sum of four terms, what do you get? Did you mean that you got ##24I+6A^2-4A^2+A^4##? In that case, there's a sign error. You can of course also simplify the two terms in the middle to ##2A^2##.

Do you get a different result for the right-hand side?
 
Last edited:
  • Like
Likes 1 person
  • #8
Fredrik said:
Oops. It's not. I didn't think it through. Sorry about that. 24AB+24BA (which appears in the left-hand side when you rewrite it as a sum of four terms) is not always equal to 48AB (which appears in the right-hand side). So we can conclude that (a) is very likely not true for all A,B. To complete the proof, you should look for a specific choice of A and B that makes the equality in (a) false.Your book should include a definition of matrix multiplication, a definition of the transpose operation, and several theorems that state elementary results like ##(A+B)C=AC+BC## and ##(A^T)^T=A##. One way to include more information is to use only one of those theorems at a time. But if you want to do things like (A+B)(C-D)=AC-AD+BC-BD in one step, that's OK too. You just need to make sure that your readers can tell what you're doing.I'm not sure sure what you're doing here. If you just take the left-hand side ##(6I+A^2)(-4I+A^2)## and rewrite it as a sum of four terms, what do you get? Did you mean that you got ##24I+6A^2-4A^2+A^4##? In that case, there's a sign error. You can of course also simplify the two terms in the middle to ##2A^2##.

Do you get a different result for the right-hand side?

Yes that's was what I meant. what do you mean by sign error?

(a) must be false. The system states so. But I do not know why it's false.

Let me rework it.

Edit: Ah I missed it. In ref to part(a), 24AB + 24 BA is not necessarily 48 AB as you've state. It could be but it is not a corollary that AB + BA = 2AB.
 
Last edited:
  • #9
Fredrik said:
Oops. It's not. I didn't think it through. Sorry about that. 24AB+24BA (which appears in the left-hand side when you rewrite it as a sum of four terms) is not always equal to 48AB (which appears in the right-hand side). So we can conclude that (a) is very likely not true for all A,B. To complete the proof, you should look for a specific choice of A and B that makes the equality in (a) false.


Your book should include a definition of matrix multiplication, a definition of the transpose operation, and several theorems that state elementary results like ##(A+B)C=AC+BC## and ##(A^T)^T=A##. One way to include more information is to use only one of those theorems at a time. But if you want to do things like (A+B)(C-D)=AC-AD+BC-BD in one step, that's OK too. You just need to make sure that your readers can tell what you're doing.


I'm not sure sure what you're doing here. If you just take the left-hand side ##(6I+A^2)(-4I+A^2)## and rewrite it as a sum of four terms, what do you get? Did you mean that you got ##24I+6A^2-4A^2+A^4##? In that case, there's a sign error. You can of course also simplify the two terms in the middle to ##2A^2##.

Do you get a different result for the right-hand side?

for part(b)

LHS = RHS
 
  • #10
Part(a) = false, part(b) = true, part(c) = true and verified true by the system.

(A(-4 I - 6 A))T = AT(-4 I - 6 A)T is true, but why?
 
  • #11
You know that ##(A(-4I-6A))^T)=(-4I-6A)^TA^T##. If you rewrite this right-hand side as a sum of two terms, and compare the result to what you get when you rewrite the right-hand side of the equality in your post as the sum of two terms, you will find that they're the same. So ##A^T## commutes with ##(-4I-6A)^T##. This is no surprise since ##A^T## commutes with ##A^T##, and ##I^T## commutes with everything.
 
  • Like
Likes 1 person
  • #12
fredrik said:
you know that ##(a(-4i-6a))^t)=(-4i-6a)^ta^t##. If you rewrite this right-hand side as a sum of two terms, and compare the result to what you get when you rewrite the right-hand side of the equality in your post as the sum of two terms, you will find that they're the same. So ##a^t## commutes with ##(-4i-6a)^t##. This is no surprise since ##a^t## commutes with ##a^t##, and ##i^t## commutes with everything.

(a(-4i-6a))t = lhs = (-4i -6a)tat
∴ lhs = (-4i -6a)tat

rhs = -4atit - 6atat = (-4it-6at)at = (-4i - 6a)tat
 
  • #13
That's good.

I missed your question about what sign error I was referring to earlier. I meant that when you evaluate ##(6I+A^2)(-4I+A^2)##, one of the terms is ##-24I##, not ##+24I##.
 
  • Like
Likes 1 person

FAQ: Difficult Matrix Identity Question

What is a difficult matrix identity?

A difficult matrix identity is a mathematical expression that relates matrices to each other in a complex way. It often involves multiple matrices and operations, making it challenging to simplify or solve.

Why are difficult matrix identities important?

Difficult matrix identities are important because they play a crucial role in many areas of mathematics, physics, and engineering. They are used to solve equations, make predictions, and analyze data.

What are some examples of difficult matrix identities?

Some examples of difficult matrix identities include the Sylvester's Law of Inertia, Cayley-Hamilton theorem, and the Kronecker product identity. These identities involve multiple matrices and require advanced algebraic manipulation to solve.

How can I solve a difficult matrix identity?

To solve a difficult matrix identity, it is helpful to first break it down into smaller, more manageable parts. Then, use mathematical properties and techniques, such as matrix multiplication and row operations, to simplify the expression. It may also be useful to use software or calculators to assist with computations.

Are difficult matrix identities used in real-world applications?

Yes, difficult matrix identities are used in many real-world applications, such as in signal processing, computer graphics, and machine learning. They are also used in solving problems in physics and engineering, such as in quantum mechanics and control systems.

Similar threads

Replies
12
Views
1K
Replies
2
Views
691
Replies
2
Views
2K
Replies
3
Views
899
Replies
8
Views
3K
Replies
13
Views
2K
Replies
25
Views
3K
Back
Top