[seqfan] Re: A055458: solution to phi(c+46)=c+46 for composite c

[israel at math.ubc.ca]

Values in these cases, if any, are > 3*10^6. Heuristically, phi(n) is 
somewhat less than n, so the probability that phi(x+2*n) = phi(x) + 2*n is 
a bit more than 1/x. On the basis of this very crude heuristic we might 
expect that the values do exist (but they could be quite large). I don't 
see any obvious obstruction (which doesn't mean there isn't any).

Evidently the data given are wrong for n >= 23: it looks like the 
program used to calculate them skipped over values that were not found, so
the values given for n=23,24,...,33 should actually be for 24,25,...,34,
those for n=34,35,...,46 should be for 36,37,...,48, etc.

Cheers,
Robert

On Apr 28 2015, Richard J. Mathar wrote:

>Does anyone have information of the values of A055458 at
>n=23, 35, 49, 59, 61,...? Are there composite c such that phi(c+46)=c+46,
>for example?
>
>RJM
>
>_______________________________________________
>
>Seqfan Mailing list - http://list.seqfan.eu/
>
>