[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference rusure::math

Title:Mathematics at DEC
Moderator:RUSURE::EDP
Created:Mon Feb 03 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2083
Total number of notes:14613

1719.0. "numerical puzzle" by HERON::BUCHANAN (The was not found.) Fri Feb 26 1993 03:38

Find natural numbers a & b satisfying:

	(i) 7 doesn't divide ab(a+b)
	(ii) 7^7 does divide (a+b)^7 - a^7 - b^7

Andrew
T.RTitleUserPersonal
Name
DateLines
1719.1SSAG::LARYLaughter & hope & a sock in the eyeSat Feb 27 1993 03:0020
With a little algebra, the answer appears to be equivalent to finding a,b such
that 

		7 does not divide ab(a+b)

		7^3 divides a^2 + ab + b^2

and this sets some constraints on a and b, namely that they must be congruent
to one of the following pairs of numbers mod 7:

		a	b
	       ---     ---
		1	2
		1	4
		2	4
		3	5
		3	6
		5	6

But I'm going skiing in the morning and I can't go on like this...
1719.2As I was dropping off...SSAG::LARYLaughter & hope & a sock in the eyeSat Feb 27 1993 03:162
		a = 1 and b = 343k + 324 is one solution...

1719.4an answerHERON::BUCHANANThe was not found.Wed Mar 03 1993 04:0712
It's simpler than I'd thought.

	(a+b)^7 - a^7 - b^7 = 7ab(a�+ab+b�)�

Therefore:
	7^7 | (a+b)^7 - a^7 - b^7
=>	7^3 | a� + ab + b�
=>	7^3 | a� - b�
=>	7^3 | (a/b)� - 1

=>	a/b == 18 or (18� =) -19 mod 343
and that's it.
1719.5STAR::ABBASIi think iam psychicWed Mar 03 1993 08:0717
ref .4
        
>	(a+b)^7 - a^7 - b^7 = 7ab(a�+ab+b�)�
    
    humm? that can't be, 'cause the right hand side will generate a term
    with at most a^5 and b^5 in them.
    while the left hand side will have an a^6 and b^6 terms in them!
    
    did i , did i, finnd a mistake by Andrew!
    
    if so, i want to save this note and decorate too.
    
    \bye
    \nasser
     ps. in case iam wrong, please note i have not had my morning coffe yet.
    
    
1719.6fooeyHERON::BUCHANANThe was not found.Wed Mar 03 1993 08:4318
>>	(a+b)^7 - a^7 - b^7 = 7ab(a�+ab+b�)�
>    humm? that can't be

Oops.

maple
factor ((a+b)^7 - a^7 - b^7 );

                2          2 2
7 a b (a + b) (a  + a b + b )

Drat, I forgot to remember one factor.   It doesn't affect the proof, though.
    
>    did i , did i, finnd a mistake by Andrew!

Thanks for the flattery, but my errors are all too common :-)

Andrew.
1719.7AUSSIE::GARSONWed Mar 03 1993 20:539
re .4
    
>	=>	7^3 | (a/b)� - 1
>
>	=>	a/b == 18 or (18� =) -19 mod 343
    
    Is there any easy way to get from 7� | x� - 1 to x = 18 as one solution?
    
    Also, you don't mention why you eliminate the case a/b == 1 (mod 343).
1719.8SSAG::LARYLaughter & hope & a sock in the eyeThu Mar 04 1993 14:0919
Re .-1: In the derivation from .4,
...
=>	7^3 | a� + ab + b�
=>	7^3 | a� - b�
=>	7^3 | (a/b)� - 1

That second step isn't reversible if (a-b) is divisible by 7. That's why
a = b (mod 343) isn't a solution, and neither are some of the cases where
a = 18b (mod 343) and a = -19b (mod 343). For example, (a,b) = (7,126) is
obviously not a solution since 7 | ab(a+b). Its simple to show that if
a = b (mod 7) then, given the other constraints, a = 0 (mod 7) and its not a
solution.

So, the complete set of solutions therefore is given by:

a = 18b (mod 343) or a = -19b (mod 343), AND a <> b (mod 7)

(As for extracting the cube roots of 1 (mod 343), I dunno, but an exhaustive
search takes minutes to code and milliseconds to run...) 
1719.9HERON::BUCHANANThe was not found.Fri Mar 05 1993 03:486
The point of bringing in the (a/b)�-1 business is to show why there
are two solutions and each is the square of the other, mod 343.   When it comes
to *finding* the solutions, I spose you can take a�+ab+b� and complete the
square mod 343.   This just leaves a square root mod 343 to find.   By machine
I'd numbercrunch, by hand I'd solve mod 7, then mod 49, finally mod 343.
Andrew.