OFFSET
1,2
COMMENTS
If making a statistical bar chart with x-axis denoting the number of decompositions of an even number, and y-axis denoting the number of hits of an x-axis value for all 3 <= k <= m, there are one or more commonest x value presenting with highest y value. Such commonest x values increase when m increases, and fall on the x values listed in this sequence.
Hypothesis: With the increase of m, the commonest number of decompositions of 2n into an unordered sum of two odd primes in the range of 3 <= k <= m ascends.
This hypothesis derives that the corresponding smallest m to the terms of this sequence makes an ascending sequence. Or say, when testing with m ascending, once a number a(n) enters this sequence, no number smaller than a(n) will be able to enter this sequence if they had not enter previous.
LINKS
Lei Zhou, Table of n, a(n) for n = 1..45
EXAMPLE
When m=3, k has only one value 3, 2k=6=3+3. Only one possible decomposition, making a decomposition statistics {{x,y}}={{1,1}}. So a(1)=1;
When m=4, k gets another value 4, 2k=8=3+5. The decomposition statistics {{x,y}}={{1,2}};...
Thereafter, k=5 makes 2k=10=5+5=3+7, {{x,y}}={{1,2},{2,1}}, the commonest value is still 1.
k=6, 2k=12=5+7, {{x,y}}={{1,3},{2,1}}, commonest x is still 1.
k=7, 2k=14=3+11=7+7, {{x,y}}={{1,3},{2,2}}, commonest x is still 1.
k=8, 2k=16=3+13=5+11, {{x,y}}={{1,3},{2,3}}, except for 1, 2 is now eligible to be the new possible commonest x, so a(2)=2.
...
Counting up to k=28, the decomposition statistics is {{1,3},{2,8},{3,8},{4,5},{5,2}}, 2 and 3 are now the commonest decompositions. It is the first time for 3 to appear. So a(3)=3.
MATHEMATICA
check=0; posts={}; mpos=0; res={}; sres=0; s={}; size=0; k=2;
While[k++; k2=2*k; p2=k-1; ct=0;
While[p2=NextPrime[p2]; p2<k2, p1=k2-p2; If[PrimeQ[p1], ct++]];
(*Calculate Goldbach decomposition*)
If[ct>size, Do[AppendTo[s, 0], {i, size+1, ct}]; size=ct];
(*and construct statistics in array s*)
s[[ct]]++; m=Max[s]; aa=Position[s, m]; la=Length[aa];
Do[a=aa[[pos, 1]];
If[a>sres,
While[sres<a, AppendTo[res, 0]; sres++]; res[[a]]=n; goal=Length[res];
(*Generate list of n values where a new commonest appears*)
If[mpos<goal, mpos=goal; check++; AppendTo[posts, mpos]]],
(*Compose elements of this sequence into a list*)
{pos, 1, la}];
check<16];
posts
CROSSREFS
KEYWORD
nonn,hard
AUTHOR
Lei Zhou, Oct 29 2013
EXTENSIONS
Lei Zhou, Nov 08 2013, uploaded a b-file extending the known elements of this sequence to the 45th.
STATUS
approved