Year-end appeal: Please make a donation to the OEIS Foundation to support ongoing development and maintenance of the OEIS. We are now in our 61st year, we have over 378,000 sequences, and we’ve reached 11,000 citations (which often say “discovered thanks to the OEIS”).
%I #30 Oct 13 2015 13:51:24
%S 6,27,167,1206,9442,77563,658097,5714972,50503822,452425909,
%T 4097411586,37441633014,344698955565,3193520274110,29747746198318,
%U 278407464679282,2616351626277085,24676888631241563,233501199663256017,2215874110986269907
%N Floor( 10^n / sum(k=3..10^n, 1/k ) ).
%C n/(Sum_{k=3..n} 1/k) is a better approximation to pi(n) than Gauss' Li(n) for 15 < n < 2803.
%D Wacław Sierpiński, Co wiemy, a czego nie wiemy o liczbach pierwszych. Warsaw: PZWS, 1961, p. 21.
%H Arkadiusz Wesolowski, <a href="/A202766/b202766.txt">Table of n, a(n) for n = 1..200</a>
%H Eric Weisstein's World of Mathematics, <a href="http://mathworld.wolfram.com/PrimeCountingFunction.html">Prime Counting Function</a>
%H Eric Weisstein's World of Mathematics, <a href="http://mathworld.wolfram.com/PrimeNumberTheorem.html">Prime Number Theorem</a>
%F a(n) = floor((10^n)/(Sum_{k=3..10^n} 1/k)).
%F a(n) ~ 10^n/(log(10^n) + gamma - 3/2).
%e a(2) = 27 because (10^2)/(Sum_{k=3..100} 1/k) = 27.1195448585....
%t lst = {}; Do[AppendTo[lst, Floor[10^n/(NIntegrate[(1 - x^10^n)/(1 - x), {x, 0, 1}, WorkingPrecision -> 20] - 1.5)]], {n, 13}]; lst
%Y Cf. A000720, A006880, A193257.
%K nonn
%O 1,1
%A _Arkadiusz Wesolowski_, Dec 23 2011