Is $prod_1leq i< jleq n fraca_i - a_ji-j$, with distinct integers $a_i$, an integer?
Clash Royale CLAN TAG#URR8PPP
up vote
16
down vote
favorite
It is known that for every $n$ consecutive integers, their product is divisible by $n!$, since $$mchoosen = fracm!n!(m-n)!$$ is also an integer.
So is it true that for every distinct integer $a_1, a_2, ..., a_n$, the number $$S = prod_1leq i< jleq n fraca_i - a_ji-j$$ is also an integer?
(Sorry for my grammar mistakes, English is my second language)
combinatorics algebra-precalculus geometry number-theory
add a comment |Â
up vote
16
down vote
favorite
It is known that for every $n$ consecutive integers, their product is divisible by $n!$, since $$mchoosen = fracm!n!(m-n)!$$ is also an integer.
So is it true that for every distinct integer $a_1, a_2, ..., a_n$, the number $$S = prod_1leq i< jleq n fraca_i - a_ji-j$$ is also an integer?
(Sorry for my grammar mistakes, English is my second language)
combinatorics algebra-precalculus geometry number-theory
What will happen If each $a_i$ is a multiple of some prime $p>n$?
â Jens Schwaiger
Aug 30 at 5:54
1
@JensSchwaiger I think the problem is still true i guess? I have tried with 4,5,6,7 and it satisfied that $S$ was an integer.
â apple
Aug 30 at 5:57
Forget my previous comment. I mixed up nominator and denominator. Probably still tired.
â Jens Schwaiger
Aug 30 at 6:06
For what it's worth, the conjecture is true for all $nleq 7$, and I'm having a devil of a time finding counter-examples with simulations for larger $n$.
â Hans Musgrave
Aug 30 at 8:48
add a comment |Â
up vote
16
down vote
favorite
up vote
16
down vote
favorite
It is known that for every $n$ consecutive integers, their product is divisible by $n!$, since $$mchoosen = fracm!n!(m-n)!$$ is also an integer.
So is it true that for every distinct integer $a_1, a_2, ..., a_n$, the number $$S = prod_1leq i< jleq n fraca_i - a_ji-j$$ is also an integer?
(Sorry for my grammar mistakes, English is my second language)
combinatorics algebra-precalculus geometry number-theory
It is known that for every $n$ consecutive integers, their product is divisible by $n!$, since $$mchoosen = fracm!n!(m-n)!$$ is also an integer.
So is it true that for every distinct integer $a_1, a_2, ..., a_n$, the number $$S = prod_1leq i< jleq n fraca_i - a_ji-j$$ is also an integer?
(Sorry for my grammar mistakes, English is my second language)
combinatorics algebra-precalculus geometry number-theory
combinatorics algebra-precalculus geometry number-theory
edited Aug 30 at 8:16
Blue
44.1k868141
44.1k868141
asked Aug 30 at 5:30
apple
844
844
What will happen If each $a_i$ is a multiple of some prime $p>n$?
â Jens Schwaiger
Aug 30 at 5:54
1
@JensSchwaiger I think the problem is still true i guess? I have tried with 4,5,6,7 and it satisfied that $S$ was an integer.
â apple
Aug 30 at 5:57
Forget my previous comment. I mixed up nominator and denominator. Probably still tired.
â Jens Schwaiger
Aug 30 at 6:06
For what it's worth, the conjecture is true for all $nleq 7$, and I'm having a devil of a time finding counter-examples with simulations for larger $n$.
â Hans Musgrave
Aug 30 at 8:48
add a comment |Â
What will happen If each $a_i$ is a multiple of some prime $p>n$?
â Jens Schwaiger
Aug 30 at 5:54
1
@JensSchwaiger I think the problem is still true i guess? I have tried with 4,5,6,7 and it satisfied that $S$ was an integer.
â apple
Aug 30 at 5:57
Forget my previous comment. I mixed up nominator and denominator. Probably still tired.
â Jens Schwaiger
Aug 30 at 6:06
For what it's worth, the conjecture is true for all $nleq 7$, and I'm having a devil of a time finding counter-examples with simulations for larger $n$.
â Hans Musgrave
Aug 30 at 8:48
What will happen If each $a_i$ is a multiple of some prime $p>n$?
â Jens Schwaiger
Aug 30 at 5:54
What will happen If each $a_i$ is a multiple of some prime $p>n$?
â Jens Schwaiger
Aug 30 at 5:54
1
1
@JensSchwaiger I think the problem is still true i guess? I have tried with 4,5,6,7 and it satisfied that $S$ was an integer.
â apple
Aug 30 at 5:57
@JensSchwaiger I think the problem is still true i guess? I have tried with 4,5,6,7 and it satisfied that $S$ was an integer.
â apple
Aug 30 at 5:57
Forget my previous comment. I mixed up nominator and denominator. Probably still tired.
â Jens Schwaiger
Aug 30 at 6:06
Forget my previous comment. I mixed up nominator and denominator. Probably still tired.
â Jens Schwaiger
Aug 30 at 6:06
For what it's worth, the conjecture is true for all $nleq 7$, and I'm having a devil of a time finding counter-examples with simulations for larger $n$.
â Hans Musgrave
Aug 30 at 8:48
For what it's worth, the conjecture is true for all $nleq 7$, and I'm having a devil of a time finding counter-examples with simulations for larger $n$.
â Hans Musgrave
Aug 30 at 8:48
add a comment |Â
2 Answers
2
active
oldest
votes
up vote
14
down vote
We observe that $prod_1 leq i < j leq n (i-j) = prod_r=1^n r!$ and the numerator is just the Vandermonde Determinant so that by performing row operations we obtain
$$prod_1 leq i < j leq n fraca_i-a_ji-j =
beginvmatrix
1 & 1 & cdots & 1 \
binoma_11 & binoma_21 & cdots & binoma_n1 \
binoma_12 & binoma_22 & cdots & binoma_n2 \
cdots & cdots & cdots & cdots \
binoma_1n-1 & binoma_2n-1 & cdots & binoma_nn-1 \
endvmatrix$$
which is obviously an integer.
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
2
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
add a comment |Â
up vote
2
down vote
Let $q = p^k$ for some prime $p$ and integer $k>0$. If you let $n_q$ = the number of pairs $i<j$ for which $a_i - a_j$ is divisible by $q$, then notice that $n_q$ is minimized for $a_1,ldots,a_n = 1,2,ldots,n$ (it is a bit of work to show this but it is not deep) (basically, you consider the set $mathbbZ/(q)$ and now you look the function $f: mathbbZ/(q) rightarrow mathbbN$ for which $f(a)$ is the number of $i$ for which $a_i$ mod $q$ is $a$. Then observe that you can compute $n_q$ from $f$, and that $n_q$ is minimized if all values of $f$ are inside $m,m+1$ where $m$ is $n/q$ rounded down).
Now apply this for $q = p, p^2, p^3, ldots$ and you see that the $p$-adic valuation of the numerator of $S$ is at least that of the denominator. That means $S$ is an integer.
add a comment |Â
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
up vote
14
down vote
We observe that $prod_1 leq i < j leq n (i-j) = prod_r=1^n r!$ and the numerator is just the Vandermonde Determinant so that by performing row operations we obtain
$$prod_1 leq i < j leq n fraca_i-a_ji-j =
beginvmatrix
1 & 1 & cdots & 1 \
binoma_11 & binoma_21 & cdots & binoma_n1 \
binoma_12 & binoma_22 & cdots & binoma_n2 \
cdots & cdots & cdots & cdots \
binoma_1n-1 & binoma_2n-1 & cdots & binoma_nn-1 \
endvmatrix$$
which is obviously an integer.
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
2
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
add a comment |Â
up vote
14
down vote
We observe that $prod_1 leq i < j leq n (i-j) = prod_r=1^n r!$ and the numerator is just the Vandermonde Determinant so that by performing row operations we obtain
$$prod_1 leq i < j leq n fraca_i-a_ji-j =
beginvmatrix
1 & 1 & cdots & 1 \
binoma_11 & binoma_21 & cdots & binoma_n1 \
binoma_12 & binoma_22 & cdots & binoma_n2 \
cdots & cdots & cdots & cdots \
binoma_1n-1 & binoma_2n-1 & cdots & binoma_nn-1 \
endvmatrix$$
which is obviously an integer.
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
2
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
add a comment |Â
up vote
14
down vote
up vote
14
down vote
We observe that $prod_1 leq i < j leq n (i-j) = prod_r=1^n r!$ and the numerator is just the Vandermonde Determinant so that by performing row operations we obtain
$$prod_1 leq i < j leq n fraca_i-a_ji-j =
beginvmatrix
1 & 1 & cdots & 1 \
binoma_11 & binoma_21 & cdots & binoma_n1 \
binoma_12 & binoma_22 & cdots & binoma_n2 \
cdots & cdots & cdots & cdots \
binoma_1n-1 & binoma_2n-1 & cdots & binoma_nn-1 \
endvmatrix$$
which is obviously an integer.
We observe that $prod_1 leq i < j leq n (i-j) = prod_r=1^n r!$ and the numerator is just the Vandermonde Determinant so that by performing row operations we obtain
$$prod_1 leq i < j leq n fraca_i-a_ji-j =
beginvmatrix
1 & 1 & cdots & 1 \
binoma_11 & binoma_21 & cdots & binoma_n1 \
binoma_12 & binoma_22 & cdots & binoma_n2 \
cdots & cdots & cdots & cdots \
binoma_1n-1 & binoma_2n-1 & cdots & binoma_nn-1 \
endvmatrix$$
which is obviously an integer.
answered Aug 30 at 13:33
A. S. Roy
1412
1412
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
2
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
add a comment |Â
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
2
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I don't see how the first relation holds: $prod_1 leq i < j leq 2 (i - j) = (1 - 2) = -1 neq 1! times 2!$
â Arin Chaudhuri
Aug 30 at 14:59
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
I think that $prod_r=1^n r!$ should be $pm prod_r=1^n-1 r!$. But apart from that, the proof works. Notice how on the $r+1$'th row you have $a_1^r/r!$ plus lower powers of $a_1$, but those lower powers can be row-reduced away using previous rows. This means that after row-reduction you get a Vandermonde matrix with the $r+1$'th row divided by $r!$. So you get the determinant of the Vandermonde matrix, divided by $prod_r=1^n-1 r!$. With that being an integer, the result follows.
â Mark
Aug 30 at 15:06
2
2
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
I believe this is what people call a proof from the book. +1.
â J.G.
Aug 30 at 15:21
add a comment |Â
up vote
2
down vote
Let $q = p^k$ for some prime $p$ and integer $k>0$. If you let $n_q$ = the number of pairs $i<j$ for which $a_i - a_j$ is divisible by $q$, then notice that $n_q$ is minimized for $a_1,ldots,a_n = 1,2,ldots,n$ (it is a bit of work to show this but it is not deep) (basically, you consider the set $mathbbZ/(q)$ and now you look the function $f: mathbbZ/(q) rightarrow mathbbN$ for which $f(a)$ is the number of $i$ for which $a_i$ mod $q$ is $a$. Then observe that you can compute $n_q$ from $f$, and that $n_q$ is minimized if all values of $f$ are inside $m,m+1$ where $m$ is $n/q$ rounded down).
Now apply this for $q = p, p^2, p^3, ldots$ and you see that the $p$-adic valuation of the numerator of $S$ is at least that of the denominator. That means $S$ is an integer.
add a comment |Â
up vote
2
down vote
Let $q = p^k$ for some prime $p$ and integer $k>0$. If you let $n_q$ = the number of pairs $i<j$ for which $a_i - a_j$ is divisible by $q$, then notice that $n_q$ is minimized for $a_1,ldots,a_n = 1,2,ldots,n$ (it is a bit of work to show this but it is not deep) (basically, you consider the set $mathbbZ/(q)$ and now you look the function $f: mathbbZ/(q) rightarrow mathbbN$ for which $f(a)$ is the number of $i$ for which $a_i$ mod $q$ is $a$. Then observe that you can compute $n_q$ from $f$, and that $n_q$ is minimized if all values of $f$ are inside $m,m+1$ where $m$ is $n/q$ rounded down).
Now apply this for $q = p, p^2, p^3, ldots$ and you see that the $p$-adic valuation of the numerator of $S$ is at least that of the denominator. That means $S$ is an integer.
add a comment |Â
up vote
2
down vote
up vote
2
down vote
Let $q = p^k$ for some prime $p$ and integer $k>0$. If you let $n_q$ = the number of pairs $i<j$ for which $a_i - a_j$ is divisible by $q$, then notice that $n_q$ is minimized for $a_1,ldots,a_n = 1,2,ldots,n$ (it is a bit of work to show this but it is not deep) (basically, you consider the set $mathbbZ/(q)$ and now you look the function $f: mathbbZ/(q) rightarrow mathbbN$ for which $f(a)$ is the number of $i$ for which $a_i$ mod $q$ is $a$. Then observe that you can compute $n_q$ from $f$, and that $n_q$ is minimized if all values of $f$ are inside $m,m+1$ where $m$ is $n/q$ rounded down).
Now apply this for $q = p, p^2, p^3, ldots$ and you see that the $p$-adic valuation of the numerator of $S$ is at least that of the denominator. That means $S$ is an integer.
Let $q = p^k$ for some prime $p$ and integer $k>0$. If you let $n_q$ = the number of pairs $i<j$ for which $a_i - a_j$ is divisible by $q$, then notice that $n_q$ is minimized for $a_1,ldots,a_n = 1,2,ldots,n$ (it is a bit of work to show this but it is not deep) (basically, you consider the set $mathbbZ/(q)$ and now you look the function $f: mathbbZ/(q) rightarrow mathbbN$ for which $f(a)$ is the number of $i$ for which $a_i$ mod $q$ is $a$. Then observe that you can compute $n_q$ from $f$, and that $n_q$ is minimized if all values of $f$ are inside $m,m+1$ where $m$ is $n/q$ rounded down).
Now apply this for $q = p, p^2, p^3, ldots$ and you see that the $p$-adic valuation of the numerator of $S$ is at least that of the denominator. That means $S$ is an integer.
edited Aug 30 at 13:34
answered Aug 30 at 13:27
Mark
1,00657
1,00657
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%2f2899169%2fis-prod-1-leq-i-j-leq-n-fraca-i-a-ji-j-with-distinct-integers-a-i%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
What will happen If each $a_i$ is a multiple of some prime $p>n$?
â Jens Schwaiger
Aug 30 at 5:54
1
@JensSchwaiger I think the problem is still true i guess? I have tried with 4,5,6,7 and it satisfied that $S$ was an integer.
â apple
Aug 30 at 5:57
Forget my previous comment. I mixed up nominator and denominator. Probably still tired.
â Jens Schwaiger
Aug 30 at 6:06
For what it's worth, the conjecture is true for all $nleq 7$, and I'm having a devil of a time finding counter-examples with simulations for larger $n$.
â Hans Musgrave
Aug 30 at 8:48