cp's OEIS Frontend

This is a front-end for the Online Encyclopedia of Integer Sequences, made by Christian Perfect. The idea is to provide OEIS entries in non-ancient HTML, and then to think about how they're presented visually. The source code is on GitHub.

Showing 1-1 of 1 results.

A278638 Numbers n such that 1/n is a difference of Egyptian fractions with all denominators < n.

Original entry on oeis.org

6, 12, 15, 18, 20, 21, 24, 28, 30, 33, 35, 36, 40, 42, 44, 45, 48, 52, 54, 55, 56, 60, 63, 65, 66, 68, 70, 72, 75, 76, 77, 78, 80, 84, 85, 88, 90, 91, 95, 96, 99, 100, 102, 104, 105, 108, 110, 112, 114, 115, 117, 119, 120, 126, 130, 132, 133, 135, 136, 138, 140, 143, 144, 145, 147, 150, 152, 153
Offset: 1

Views

Author

Robert Israel, Nov 24 2016

Keywords

Comments

Numbers n such that we can write 1/n = Sum_{1<=k
Numbers n such that A072207(n) < 2*A072207(n-1).
If n is in the sequence, so is k*n for all k>1 (cf. A328226).
Contains A001284, because 1/(m*k) = 1/(m*(k-m))-1/(k*(k-m)).
Disjoint from A000961.
2*p^k with p prime is in the sequence if and only if p=3.
3*p^k with p prime is in the sequence if and only if p=2,5,7 or 11.
4*p^k with p prime is in the sequence if and only if p=3,5,7,11,13,17 or 19.
For each m that is not a term, there are only finitely many primes p such that some m*p^k is a term. [Corrected by Max Alekseyev, Oct 08 2019]

Examples

			44 is in the sequence because 1/44 = (1/12 + 1/33) - 1/11.
4 is not in the sequence because 1/4 can't be written as the difference of sums of two subsets of {1, 1/2, 1/3}.
		

Crossrefs

Contains A005279. - Robert G. Wilson v, Nov 27 2016

Programs

  • Maple
    N:= 200: # to get all terms <= N
    V:= Vector(N):
    f:= proc(n)  option remember;
    local F,E,p,e,k,m,L,L1,i,s,t,sg,Maybe;
    global Rep;
      F:= numtheory:-factorset(n);
      if nops(F) = 1 then return false fi;
      if ormap(m -> n < m^2 and m^2 < 2*n, numtheory:-divisors(n)) then
        for m in numtheory:-divisors(n) do
          if n < m^2 and m^2 < 2*n then
            k:= n/m; Rep[n]:= [m*(k-m),-k*(k-m)]; return true
          fi
        od
      fi;
      F:= convert(F,list);
      E:= map(p -> padic:-ordp(n,p), F);
      i:= max[index](zip(`^`,F,E));
      p:= F[i];
      e:= E[i];
      k:= n/p^e;
      Maybe:= false;
      for i from 3^(k-1) to 2*3^(k-1)-1 do
        L:= (-1) +~ convert(i,base,3);
        s:= 1/k - add(L[i]/i,i=1..k-1);
        if numer(s) mod p = 0 then
        Maybe:= true;
          t:= abs(s/p^e); sg:= signum(s);
          if  (numer(t) <= 1 and (denom(t) < n or (denom(t) < N and V[denom(t)] = 1))) or (numer(t) = 2 and denom(t) < N and V[denom(t)] = 1) then
             L1:= subs(0=NULL, [seq(L[i]*i*p^e,i=1..k-1)]);
             if t = 0 then ;
             elif numer(t) = 1 and denom(t) < n then L1:= [op(L1),sg/t]
             elif procname(2/t) then
                L1:= ([op(L1), 2*sg/t, op(expand(sg*Rep[2/t]))])
             else next
             fi;
             if max(abs~(L1)) < n then Rep[n]:= L1; return true fi;
          fi;
        fi
      od:
      if Maybe then printf("Warning: %d is uncertain\n",n)
    else false
    fi;
    end proc:
    for n from 6 to N do
      if V[n] = 0 and f(n) then
        V[n] := 1;
        for j from 2*n to N by n do
          if not assigned(Rep[j]) then
            V[j]:= 1;
            Rep[j] := map(`*`,Rep[n],j/n);
            f(j):= true;
          fi
        od;
      fi;
    od:
    select(t -> V[t]=1,[$6..N]);
  • Mathematica
    sol[n_] := Module[{c, cc}, cc = Array[c, n-1]; FindInstance[AllTrue[cc, -1 <= # <= 1&] && 1/n == Total[cc/Range[n-1]], cc, Integers, 1]];
    Reap[For[n = 6, n <= 200, n++, If[sol[n] != {}, Print[n]; Sow[n]]]][[2, 1]] (* Jean-François Alcover, Jul 29 2020 *)
Showing 1-1 of 1 results.