Putnam and Beyond: Problem 87 (Matrices and algebra)
Clash Royale CLAN TAG#URR8PPP
up vote
0
down vote
favorite
Below is the problem as stated in Putnam and Beyond:
Let $A$ and $B$ be tw0 $ntimes n$ matrices that commute and such that for some positive integers $p$ and $q$, $A^p=I_n$ and $B^q=O_n$. Prove that $A+B$ is invertible, and find its inverse.
In the following I will give my "solution" and I would be very grateful if you could either verify that it is correct or else tell me where I've gone wrong. I would also greatly appreciate any comments on my proof writing.
We start of by making the assumption $p>q$. Note that in order for this argument to be correct, $p$ and $q$ may need to be interchanged. It follows that $A^p+B^p=I_n$. The expression on the left can be factored as $(A+B)(A^p-1+BA^p-2+...+B^p-1)= I_n$. From this we see clearly that $A+B$ is invertible, its inverse being $(A^p-1+BA^p-2+...+B^p-1)$. My argument is concluded.
Thank you!
proof-verification proof-writing contest-math
 |Â
show 4 more comments
up vote
0
down vote
favorite
Below is the problem as stated in Putnam and Beyond:
Let $A$ and $B$ be tw0 $ntimes n$ matrices that commute and such that for some positive integers $p$ and $q$, $A^p=I_n$ and $B^q=O_n$. Prove that $A+B$ is invertible, and find its inverse.
In the following I will give my "solution" and I would be very grateful if you could either verify that it is correct or else tell me where I've gone wrong. I would also greatly appreciate any comments on my proof writing.
We start of by making the assumption $p>q$. Note that in order for this argument to be correct, $p$ and $q$ may need to be interchanged. It follows that $A^p+B^p=I_n$. The expression on the left can be factored as $(A+B)(A^p-1+BA^p-2+...+B^p-1)= I_n$. From this we see clearly that $A+B$ is invertible, its inverse being $(A^p-1+BA^p-2+...+B^p-1)$. My argument is concluded.
Thank you!
proof-verification proof-writing contest-math
looks fine to me. You can probably do a little more and say why you can assume $p>q$ in the first place.
â dezdichado
Aug 9 at 18:29
@RobertIsrael OK, so how about using A^p-(-B)^p =(A+B)(A^p-1 .... + B^p-1). I don't see the need of introducing k, can you explain?
â wittbluenote
Aug 9 at 18:41
Actually what you want is $A^kp + (-1)^kp B^kp = I_n$ where $kp > q$. BTW your factoring is wrong.
â Robert Israel
Aug 9 at 18:42
The point is that in the case $p < q$, $B^p$ isn't $O$, while $A^q$ isn't $I$.
â Robert Israel
Aug 9 at 18:44
@RobertIsrael In the case $p<q$ isn't simply interchanging p and q in my argument sufficient. That is instead of considering $A^p + B^p$, we consider $A^q+ B^q$. ( That is what I meant by:" Note that in order for this argument to be correct, p and q may need to be interchanged")
â wittbluenote
Aug 9 at 18:46
 |Â
show 4 more comments
up vote
0
down vote
favorite
up vote
0
down vote
favorite
Below is the problem as stated in Putnam and Beyond:
Let $A$ and $B$ be tw0 $ntimes n$ matrices that commute and such that for some positive integers $p$ and $q$, $A^p=I_n$ and $B^q=O_n$. Prove that $A+B$ is invertible, and find its inverse.
In the following I will give my "solution" and I would be very grateful if you could either verify that it is correct or else tell me where I've gone wrong. I would also greatly appreciate any comments on my proof writing.
We start of by making the assumption $p>q$. Note that in order for this argument to be correct, $p$ and $q$ may need to be interchanged. It follows that $A^p+B^p=I_n$. The expression on the left can be factored as $(A+B)(A^p-1+BA^p-2+...+B^p-1)= I_n$. From this we see clearly that $A+B$ is invertible, its inverse being $(A^p-1+BA^p-2+...+B^p-1)$. My argument is concluded.
Thank you!
proof-verification proof-writing contest-math
Below is the problem as stated in Putnam and Beyond:
Let $A$ and $B$ be tw0 $ntimes n$ matrices that commute and such that for some positive integers $p$ and $q$, $A^p=I_n$ and $B^q=O_n$. Prove that $A+B$ is invertible, and find its inverse.
In the following I will give my "solution" and I would be very grateful if you could either verify that it is correct or else tell me where I've gone wrong. I would also greatly appreciate any comments on my proof writing.
We start of by making the assumption $p>q$. Note that in order for this argument to be correct, $p$ and $q$ may need to be interchanged. It follows that $A^p+B^p=I_n$. The expression on the left can be factored as $(A+B)(A^p-1+BA^p-2+...+B^p-1)= I_n$. From this we see clearly that $A+B$ is invertible, its inverse being $(A^p-1+BA^p-2+...+B^p-1)$. My argument is concluded.
Thank you!
proof-verification proof-writing contest-math
edited Aug 9 at 18:27
asked Aug 9 at 18:22
wittbluenote
958
958
looks fine to me. You can probably do a little more and say why you can assume $p>q$ in the first place.
â dezdichado
Aug 9 at 18:29
@RobertIsrael OK, so how about using A^p-(-B)^p =(A+B)(A^p-1 .... + B^p-1). I don't see the need of introducing k, can you explain?
â wittbluenote
Aug 9 at 18:41
Actually what you want is $A^kp + (-1)^kp B^kp = I_n$ where $kp > q$. BTW your factoring is wrong.
â Robert Israel
Aug 9 at 18:42
The point is that in the case $p < q$, $B^p$ isn't $O$, while $A^q$ isn't $I$.
â Robert Israel
Aug 9 at 18:44
@RobertIsrael In the case $p<q$ isn't simply interchanging p and q in my argument sufficient. That is instead of considering $A^p + B^p$, we consider $A^q+ B^q$. ( That is what I meant by:" Note that in order for this argument to be correct, p and q may need to be interchanged")
â wittbluenote
Aug 9 at 18:46
 |Â
show 4 more comments
looks fine to me. You can probably do a little more and say why you can assume $p>q$ in the first place.
â dezdichado
Aug 9 at 18:29
@RobertIsrael OK, so how about using A^p-(-B)^p =(A+B)(A^p-1 .... + B^p-1). I don't see the need of introducing k, can you explain?
â wittbluenote
Aug 9 at 18:41
Actually what you want is $A^kp + (-1)^kp B^kp = I_n$ where $kp > q$. BTW your factoring is wrong.
â Robert Israel
Aug 9 at 18:42
The point is that in the case $p < q$, $B^p$ isn't $O$, while $A^q$ isn't $I$.
â Robert Israel
Aug 9 at 18:44
@RobertIsrael In the case $p<q$ isn't simply interchanging p and q in my argument sufficient. That is instead of considering $A^p + B^p$, we consider $A^q+ B^q$. ( That is what I meant by:" Note that in order for this argument to be correct, p and q may need to be interchanged")
â wittbluenote
Aug 9 at 18:46
looks fine to me. You can probably do a little more and say why you can assume $p>q$ in the first place.
â dezdichado
Aug 9 at 18:29
looks fine to me. You can probably do a little more and say why you can assume $p>q$ in the first place.
â dezdichado
Aug 9 at 18:29
@RobertIsrael OK, so how about using A^p-(-B)^p =(A+B)(A^p-1 .... + B^p-1). I don't see the need of introducing k, can you explain?
â wittbluenote
Aug 9 at 18:41
@RobertIsrael OK, so how about using A^p-(-B)^p =(A+B)(A^p-1 .... + B^p-1). I don't see the need of introducing k, can you explain?
â wittbluenote
Aug 9 at 18:41
Actually what you want is $A^kp + (-1)^kp B^kp = I_n$ where $kp > q$. BTW your factoring is wrong.
â Robert Israel
Aug 9 at 18:42
Actually what you want is $A^kp + (-1)^kp B^kp = I_n$ where $kp > q$. BTW your factoring is wrong.
â Robert Israel
Aug 9 at 18:42
The point is that in the case $p < q$, $B^p$ isn't $O$, while $A^q$ isn't $I$.
â Robert Israel
Aug 9 at 18:44
The point is that in the case $p < q$, $B^p$ isn't $O$, while $A^q$ isn't $I$.
â Robert Israel
Aug 9 at 18:44
@RobertIsrael In the case $p<q$ isn't simply interchanging p and q in my argument sufficient. That is instead of considering $A^p + B^p$, we consider $A^q+ B^q$. ( That is what I meant by:" Note that in order for this argument to be correct, p and q may need to be interchanged")
â wittbluenote
Aug 9 at 18:46
@RobertIsrael In the case $p<q$ isn't simply interchanging p and q in my argument sufficient. That is instead of considering $A^p + B^p$, we consider $A^q+ B^q$. ( That is what I meant by:" Note that in order for this argument to be correct, p and q may need to be interchanged")
â wittbluenote
Aug 9 at 18:46
 |Â
show 4 more comments
2 Answers
2
active
oldest
votes
up vote
1
down vote
accepted
Your argument is right when $pge q$ because $$B^q=0to B^p=0 text and (-B)^p=0$$also we know that $A^p-(-B)^p$ is divisible by $A-(-B)=A+B$ so your conclusion is right in this case. Now let $p<q$. Also since all the eigenvalues of $A^p$ are $1$ so all the eigenvalues of $A$ are roots of $1$ and non-zero and $A$ is invertible which means that$$A^q=A^q-p$$therefore $$A^q-p=A^q-(-B)^q=(A+B)(A^q-1-cdots pm B^q-1)$$therefore $$(A+B)(A^q-1-cdots pm B^q-1)(A^q-p)^-1=I$$or $$(A+B)^-1=left((A^q-1-cdots pm B^q-1)(A^q-p)^-1right)^-1=A^q-p(A^q-1-cdots pm B^q-1)^-1$$
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
add a comment |Â
up vote
1
down vote
Here is marginally different way, note that $A+B = A(I+ A^-1 B)$. Since $A,B$ commute, so do $A^-1, B$ and so $(A^-1B)^k = A^-kB^k$, and so, with $C= -A^-1B$, we have $C^q = 0$.
Since all of $C$'s eigenvalues are $0$, we see that $I-C$ is invertible and it is easy to check that $(I-C)(I+C+C^2+cdots+C^q-1) = I$.
Hence $(A+B)^-1 = (I-C)^-1 A^-1 = (I+C+C^2+cdots+C^q-1) A^-1$.
Since $A^p = I$, we have $A^-k = A^p-k$ and so
$(A+B)^-1 = sum_k=0^q-1 (-B)^k A^p-k$.
add a comment |Â
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
up vote
1
down vote
accepted
Your argument is right when $pge q$ because $$B^q=0to B^p=0 text and (-B)^p=0$$also we know that $A^p-(-B)^p$ is divisible by $A-(-B)=A+B$ so your conclusion is right in this case. Now let $p<q$. Also since all the eigenvalues of $A^p$ are $1$ so all the eigenvalues of $A$ are roots of $1$ and non-zero and $A$ is invertible which means that$$A^q=A^q-p$$therefore $$A^q-p=A^q-(-B)^q=(A+B)(A^q-1-cdots pm B^q-1)$$therefore $$(A+B)(A^q-1-cdots pm B^q-1)(A^q-p)^-1=I$$or $$(A+B)^-1=left((A^q-1-cdots pm B^q-1)(A^q-p)^-1right)^-1=A^q-p(A^q-1-cdots pm B^q-1)^-1$$
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
add a comment |Â
up vote
1
down vote
accepted
Your argument is right when $pge q$ because $$B^q=0to B^p=0 text and (-B)^p=0$$also we know that $A^p-(-B)^p$ is divisible by $A-(-B)=A+B$ so your conclusion is right in this case. Now let $p<q$. Also since all the eigenvalues of $A^p$ are $1$ so all the eigenvalues of $A$ are roots of $1$ and non-zero and $A$ is invertible which means that$$A^q=A^q-p$$therefore $$A^q-p=A^q-(-B)^q=(A+B)(A^q-1-cdots pm B^q-1)$$therefore $$(A+B)(A^q-1-cdots pm B^q-1)(A^q-p)^-1=I$$or $$(A+B)^-1=left((A^q-1-cdots pm B^q-1)(A^q-p)^-1right)^-1=A^q-p(A^q-1-cdots pm B^q-1)^-1$$
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
add a comment |Â
up vote
1
down vote
accepted
up vote
1
down vote
accepted
Your argument is right when $pge q$ because $$B^q=0to B^p=0 text and (-B)^p=0$$also we know that $A^p-(-B)^p$ is divisible by $A-(-B)=A+B$ so your conclusion is right in this case. Now let $p<q$. Also since all the eigenvalues of $A^p$ are $1$ so all the eigenvalues of $A$ are roots of $1$ and non-zero and $A$ is invertible which means that$$A^q=A^q-p$$therefore $$A^q-p=A^q-(-B)^q=(A+B)(A^q-1-cdots pm B^q-1)$$therefore $$(A+B)(A^q-1-cdots pm B^q-1)(A^q-p)^-1=I$$or $$(A+B)^-1=left((A^q-1-cdots pm B^q-1)(A^q-p)^-1right)^-1=A^q-p(A^q-1-cdots pm B^q-1)^-1$$
Your argument is right when $pge q$ because $$B^q=0to B^p=0 text and (-B)^p=0$$also we know that $A^p-(-B)^p$ is divisible by $A-(-B)=A+B$ so your conclusion is right in this case. Now let $p<q$. Also since all the eigenvalues of $A^p$ are $1$ so all the eigenvalues of $A$ are roots of $1$ and non-zero and $A$ is invertible which means that$$A^q=A^q-p$$therefore $$A^q-p=A^q-(-B)^q=(A+B)(A^q-1-cdots pm B^q-1)$$therefore $$(A+B)(A^q-1-cdots pm B^q-1)(A^q-p)^-1=I$$or $$(A+B)^-1=left((A^q-1-cdots pm B^q-1)(A^q-p)^-1right)^-1=A^q-p(A^q-1-cdots pm B^q-1)^-1$$
answered Aug 9 at 18:38
Mostafa Ayaz
9,1033730
9,1033730
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
add a comment |Â
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
Thank you! I think I understand the idea behind your answer, but I don't see why in the case $q>p$ considering $A^q+B^q$ and applying the same argument as in my post is insufficient.
â wittbluenote
Aug 9 at 19:00
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
You're welcome! Do you mean that $A^q+B^q=I$ still holds?
â Mostafa Ayaz
Aug 9 at 19:03
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
In the case $q>p$, yes, I would say $A^q+ B^q= I_n$ holds.
â wittbluenote
Aug 9 at 19:05
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
You're right. Then this becomes an alternative to ur argument.
â Mostafa Ayaz
Aug 9 at 19:14
add a comment |Â
up vote
1
down vote
Here is marginally different way, note that $A+B = A(I+ A^-1 B)$. Since $A,B$ commute, so do $A^-1, B$ and so $(A^-1B)^k = A^-kB^k$, and so, with $C= -A^-1B$, we have $C^q = 0$.
Since all of $C$'s eigenvalues are $0$, we see that $I-C$ is invertible and it is easy to check that $(I-C)(I+C+C^2+cdots+C^q-1) = I$.
Hence $(A+B)^-1 = (I-C)^-1 A^-1 = (I+C+C^2+cdots+C^q-1) A^-1$.
Since $A^p = I$, we have $A^-k = A^p-k$ and so
$(A+B)^-1 = sum_k=0^q-1 (-B)^k A^p-k$.
add a comment |Â
up vote
1
down vote
Here is marginally different way, note that $A+B = A(I+ A^-1 B)$. Since $A,B$ commute, so do $A^-1, B$ and so $(A^-1B)^k = A^-kB^k$, and so, with $C= -A^-1B$, we have $C^q = 0$.
Since all of $C$'s eigenvalues are $0$, we see that $I-C$ is invertible and it is easy to check that $(I-C)(I+C+C^2+cdots+C^q-1) = I$.
Hence $(A+B)^-1 = (I-C)^-1 A^-1 = (I+C+C^2+cdots+C^q-1) A^-1$.
Since $A^p = I$, we have $A^-k = A^p-k$ and so
$(A+B)^-1 = sum_k=0^q-1 (-B)^k A^p-k$.
add a comment |Â
up vote
1
down vote
up vote
1
down vote
Here is marginally different way, note that $A+B = A(I+ A^-1 B)$. Since $A,B$ commute, so do $A^-1, B$ and so $(A^-1B)^k = A^-kB^k$, and so, with $C= -A^-1B$, we have $C^q = 0$.
Since all of $C$'s eigenvalues are $0$, we see that $I-C$ is invertible and it is easy to check that $(I-C)(I+C+C^2+cdots+C^q-1) = I$.
Hence $(A+B)^-1 = (I-C)^-1 A^-1 = (I+C+C^2+cdots+C^q-1) A^-1$.
Since $A^p = I$, we have $A^-k = A^p-k$ and so
$(A+B)^-1 = sum_k=0^q-1 (-B)^k A^p-k$.
Here is marginally different way, note that $A+B = A(I+ A^-1 B)$. Since $A,B$ commute, so do $A^-1, B$ and so $(A^-1B)^k = A^-kB^k$, and so, with $C= -A^-1B$, we have $C^q = 0$.
Since all of $C$'s eigenvalues are $0$, we see that $I-C$ is invertible and it is easy to check that $(I-C)(I+C+C^2+cdots+C^q-1) = I$.
Hence $(A+B)^-1 = (I-C)^-1 A^-1 = (I+C+C^2+cdots+C^q-1) A^-1$.
Since $A^p = I$, we have $A^-k = A^p-k$ and so
$(A+B)^-1 = sum_k=0^q-1 (-B)^k A^p-k$.
answered Aug 9 at 22:48
copper.hat
123k557156
123k557156
add a comment |Â
add a comment |Â
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2877553%2fputnam-and-beyond-problem-87-matrices-and-algebra%23new-answer', 'question_page');
);
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
looks fine to me. You can probably do a little more and say why you can assume $p>q$ in the first place.
â dezdichado
Aug 9 at 18:29
@RobertIsrael OK, so how about using A^p-(-B)^p =(A+B)(A^p-1 .... + B^p-1). I don't see the need of introducing k, can you explain?
â wittbluenote
Aug 9 at 18:41
Actually what you want is $A^kp + (-1)^kp B^kp = I_n$ where $kp > q$. BTW your factoring is wrong.
â Robert Israel
Aug 9 at 18:42
The point is that in the case $p < q$, $B^p$ isn't $O$, while $A^q$ isn't $I$.
â Robert Israel
Aug 9 at 18:44
@RobertIsrael In the case $p<q$ isn't simply interchanging p and q in my argument sufficient. That is instead of considering $A^p + B^p$, we consider $A^q+ B^q$. ( That is what I meant by:" Note that in order for this argument to be correct, p and q may need to be interchanged")
â wittbluenote
Aug 9 at 18:46