[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

1867.0. "Crux Mathematicorum 1932" by RUSURE::EDP (Always mount a scratch monkey.) Tue Apr 26 1994 13:13

    Proposed by K. R. S. Sastry, Addis Ababa, Ethiopia.  (Submitted without
    solution)
    
    Trivially, if N = 1, 10, 100, ... then in each case the initial digits
    of N^3 are the digits of N.  A nontrivial example is N = 32, since 32^3
    = 32768.  Find another positive integer N with this property.
T.RTitleUserPersonal
Name
DateLines
1867.1first oneIOSG::CARLINDick Carlin IOSG, Reading, EnglandTue Apr 26 1994 14:2735
    If n^3 = n*10^m + p where p <10^m
    
    then p=qn, so n^2 = 10^m + q where q is small :-)
    
    which suggests using the manual square root method (I'm sure it's in
    here somewhere) to successively get square roots of 10,1000,100000 etc
    and round them up:
    
                  3  1  6  2  2
                 -- -- -- -- -- -- ...
                 10 00 00 00 00 00
                  9
                 --
     6(1)         1 00
                    61
                  ---- 
     62(6)          39 00
                    37 56
                    -----
     632(2)          1 44 00
                     1 26 44
                     -------
     6324(2)           17 56 00
                       12 64 84
                       --------
                       ...
    
    Trying each step we get  32^3    = 32768          good
                             317^3   = 31855013
                             3163^3  = 31644451747
                             31623^3 = 31623446801367 good
    
    and no doubt there are more
    
    Dick
1867.2More than just another one...the whole enchiladaVMSDEV::HALLYBFish have no concept of fireTue Apr 26 1994 15:406
>    ... so n^2 = 10^m + q where q is small :-)
    
    This would seem to imply an exhaustive list.
    If N� "begins with N", then N must be approximately a root of 10^K.
    
      John
1867.3RTL::GILBERTMon May 02 1994 18:0929
The non-trivial solutions for m <= 100 are:
    
    32
    31623
    316228
    3162278
    31622777
    31622776602
    316227766017
    31622776601684
    316227766016838
    3162277660168379332
    31622776601683793320
    316227766016837933200
    3162277660168379331999
    31622776601683793319989
    3162277660168379331998894
    3162277660168379331998893544433
    316227766016837933199889354443272
    3162277660168379331998893544432719
    3162277660168379331998893544432718534
    316227766016837933199889354443271853372
    3162277660168379331998893544432718533720
    31622776601683793319988935444327185337196
    316227766016837933199889354443271853371956
    316227766016837933199889354443271853371955514
    316227766016837933199889354443271853371955513933 

There appears to be a solution for each digit >= 5 in sqrt(10).
1867.4RUSURE::EDPAlways mount a scratch monkey.Tue May 09 1995 16:2861
    Solution by Peter Hurthig, Columbia College, Burnaby, B.C.
    
    We claim that: if the (p+1)st digit to the right of the decimal in
    sqrt(10) is greater than or equal to 6, then N = floor(10^p * sqrt(10))
    + 1 has the desired property.  [Solutions are shown for p = 1, 4, 5, 6,
    7, match the first five solutions in the previous response.]
    
    For N to have the desired property, N^3 = 10^m * N + a where 0 <= a <
    10^m.  Therefore
    
    		0 <= N^3 - 10^m * N < 10^m.		(1)
    
    If m = 2p, then N(N^2-10^(2p)) >= 0 implies N >= 10^p where equality
    yields the trivial solutions.  However if N = 10^p + k for some
    positive integer k then
    
    	N^3 - 10^m * N = N(N^2-10^(2p)) = (10^p + k)(2k*10^p + k^2) > 10^m,
    
    contradicting (1).  Therefore the non-trivial solutions can occur only
    if m is an odd number, m = 2p+1.  Since N^3 - 10^(2p+1) must be
    positive, N > 10^p * sqrt(10).  So if there are any solutions of (1)
    for m = 2p+1, then N = floor(10^p * sqrt(10)) + 1, the smallest integer
    which is greater than 10^p * sqrt(10), must be one of them.
    
    If we let q = 10^p * sqrt(10) and delta = N-q then N(N^2 - 10^m) < 10^m
    becomes
    
    		(q+delta)(2*q*delta + delta^2) < q^2
    
    or
    
    		(1 - 2*delta)q^2 - 3*delta^2*q - delta^3 > 0.
    
    If delta >= 1/2 then the left side is negative and there are no
    solutions.  [In particular, N = [10^p * sqrt(10)] + 1 is the only
    possible choice for N for each value of p, because any larger choice
    would mean delta >= 1. -- Ed.]  If delta < 1/2 then the inequality is
    satisfied if
                    3*delta^2 + delta*sqrt(delta^2+4*delta)
    		q > ---------------------------------------.
    	                         2 - 4*delta
    
    In particular, if delta <= 2/5 then
    
    3*delta^2 + delta*sqrt(delta^2+4*delta)    6+2*sqrt(11)
    --------------------------------------- <= ------------ < 3 < 10^p*sqrt(10)
                 2 - 4*delta                       5
    
    for all p.
    
    Therefore [since delta = N-q is the difference between 10^p * sqrt(10)
    and the next highest integer] if the (p+1)st digit after the decimal in
    the expansion of sqrt(10) is greater than or equal to 6, then
    
    		10^p * sqrt(10) - floor(10^p * sqrt(10)) = 1-delta >= 0.6,
    
    and floor(10^p * sqrt(10)) + 1 will have the desired property.
    
    It would of course be very gratifying to show that sqrt(10) is a normal
    number (i.e., all the digits occur in the expected frequency) so that
    there would be infinitely many such N.