[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

1875.0. "Crux Mathematicorum 1950" by RUSURE::EDP (Always mount a scratch monkey.) Tue May 24 1994 12:43

    Proposed by Svetlozar Doitchev, Stara Zagora, Bulgaria.
    
    The equation
    
    	2*3^x + 1 = 7*5^y,
    
    where x and y are nonnegative integers, has x=1, y=0 as one solution. 
    Find all other solutions.
T.RTitleUserPersonal
Name
DateLines
1875.1Another SolutionEMASS::ROOSSun May 29 1994 23:0611
    
    I have a solution:   x = 7 and y = 4.
    
    Since the right side must end in the digit 5, the possible values of x
    (x > 1) must be 3,7,11,15,19,...
    
    I wrote a program in Pascal, running on a VAX, that found only the
    solution above for x < 10000 and y < 7000.  I suspect this to be the
    only other solution.
    
    
1875.2RTL::GILBERTMon May 30 1994 19:584
    2*3^x + 1 = 7*5^y
    
    2*3^1 + 1 = 7*5^0 = 7
    2*3^7 + 1 = 7*5^4 = 4375
1875.3RUSURE::EDPAlways mount a scratch monkey.Tue Jun 13 1995 15:46109
    Solution by Richard K. Guy, University of Calgary, and John Selfridge,
    Northern Illinois University, DeKalb.
    
    The result can be found in a fairly famous paper [3] which deserves to
    be even more famous.  It also appears as a special case in [2]; see the
    entry 4375 in Col. 7 opposite N = 1 in Table 4 on p. 188.  For our
    solution, it is essential to know a little elementary number theory;
    there's more than enough in section 2.8 of [4].  It is also very useful
    to have access to [1] and the extension and updates circulated by
    Wagstaff.  The solution appears to demand formidable calculations, but
    there are ways to circumvent these, some of which will be indicated
    below.  See the Remark on Calculation on p. 100 of [4].
    
    If y >= 0, then x >= 1 and we can write the given equation as
    
    	6(3^(x-1) - 1) = 7(5^y - 1).					(1)
    
    If y > =, then 7 | (3^(x-1) - 1) and 6 | (x-1).
    
    The last sentence can be expressed by saying that 3 is a primitive root
    of 7.  If you were born, as one of us was, before Abel and Galois, and
    don't know any group theory, then this is a good way to begin to learn
    some.  The primitive root 3 generates the (cyclic) multiplicative group
    of nonzero residues, modulo 7:
    
    	exponent e	0 1 2 3 4 5 6 ...
    	3^e mod 7	1 3 2 6 4 5 1 ...
    
    The order of 3 and 5 (the other primitive root mod 7) is 6 (= phi(7)),
    the order of 2 and 4 is 3, that of 6 is 2, and that of 1 is 1.
    
    Lemma 2.31 of [4] states that if a has order h mod m, then the positive
    integers k such that a^k = 1 mod m are precisely those for which h | k.
    Theorem 2.40 is that if p is an odd prime and g is a primitive root
    modulo p^2, then g is a primitive root modulo p^a for a = 3, 4, 5, ...
    
    Euler's generalization of Fermat's "little" theorem states that
    
    	gcd(a,m) = 1 implies a^phi(m) = 1 mod m.
    
    [Note that phi(p^a) = p^(a-1) (p-1).]
    
    Before we interrupted ourselves, we had just seen that 6 | (x-1).  In
    fact,
    
    	x = 7, y = 4
    
    is a solution, and we will show that there are no others.  Write
    equation (1) as
    
    	2 * 3^7 (3^a-1) = 7 * 5^4 (5^b-1)				(2)
    
    where a = x-7 and b = y-4.  We will obtain a contradiction by assuming
    that a and b are positive.  Conveniently, 3 and 5 are primitive roots
    of each other and of each other's squares, and hence, by the theorem
    quoted above, of all higher powers.
    
    	exponent e	0 1 2 3 4 5 6 ...
    	5^e mod 3^2	1 5 7 8 4 2 1 ...
    
    	exponent e	0 1 2 3 4 5    10
    	3^e mod 562	1 3 9 2 6 18   24 (which is congruent to -1)
    
    (a) From (2), 5^4 || (3^a-1), so 4*5^3 | a and 5^3 || a.
    
    [We use the symbol || for "exactly divides" in the sense that 5^4
    divides 3^a-1 but 5^5 does not divided 3^a-1, so 5^3 divides a, but 5^4
    does not divide a.]
    
    (b) 3^7 || (5^b-1), so 2*3^6 | b and 3^6 || b.
    
    (c) From (a) 2^4 | (3^4-1) | (3^a-1), so 2^5 | (5^b-1) and 8 | b.
    
    (d) From (b) & (c) 24 | b, so 390001 | (5^12+1) | (5^b-1) and 390001 |
    (3^a-1) implies that 625 | a.  But this contradicts (a).
    
    End of proof.  You may have a couple of objections:  that the
    calculations in (d) are beyond the capabilities of your hand
    calculator, and where did 390001 come from?  We admit that it's very
    useful to have the tables [1].  But the calculations are easy:  the
    prime
    
    	390001 = 1 + 5^4 (5^4-1) = 1 - 5^4 + 5^8 = (5^12+1) / (5^4+1)
    
    divides 5^b-1 = 5^(24k) - 1 which is a multiple of 5^24-1 =
    (5^12-1)(5^12+1).  We also want to check that the order of 3 modulo
    390001 is divisible by 625.  Work in the scale of 625 = 5^4 (note that
    5^8 = 5^4-1 and 5^12 = -1 mod 390001):  3^6 = 5^4 +104 and we already
    know that 2*3^7 = 7*5^4-1, leading to 3^13 = 55*5^4-56 and squaring and
    multiplying (which can be done by hand calculation or a hand
    calculator) lead eventually to 3^4875 is congruent to 91*5^4-185 which
    is not congruent to 1, but whose fifth power is.
    
    
    References.
    
    [1] John Brillhart, D. H. Lehmer, J. L. Selfridge, Bryant Tuckerman &
    S. S. Wagstaff, Factorizations of b^n +/- 1, b = 2, 3, 5, 6, 7, 10, 11,
    12 up to high powers, Contemporary Math. 22, Amer. Math. Soc., Second
    edition 1988 & updating by S. S. Wagstaff.
    
    [2] R. K. Guy, C. B. Lacampagne & J. L. Selfridge, Primes at a glance,
    Math. Comput. 48 (1987) 183-202.
    
    [3] D. H. Lehmer, On a problem of Stormer, Illinois J. Math., 8 (1964)
    59-79.
    
    [4] Ivan Niven, Herbert S. Zuckerman & Hugh L. Montgomery, An
    Introduction to the Theory of Numbers, Wiley, Fifth edition, 1991.