%I A114576
%S A114576 1,1,2,3,1,6,3,11,10,23,26,2,47,70,10,102,176,45,221,449,160,5,493,1121,
%T A114576 539,35,1105,2817,1680,196,2516,7031,5082,868,14,5763,17604,14856,3486,
%U A114576 126,13328,43996,42660,12810,840,30995,110147,120338,44640,4410,42
%N A114576 Triangle read by rows: T(n,k) is number of Motzkin paths of length n 
               having k UH's, where U=(1,1), H=(1,0) (0<=k<=floor(n/3).
%C A114576 Row n contains 1+floor(n/3) terms. Row sums are the Motzkin numbers (A001006). 
               Column 0 yields A090344. Sum(k*T(n,k),k=0..floor(n/3))=A014531(n-2).
%F A114576 G.f.=[1-z-sqrt(1-2z-3z^2-4tz^3+4z^3)]/[2(1-z+tz)z^2].
%e A114576 T(4,1)=3 because we have H(UH)D, (UH)DH and (UH)HD, where U=(1,1), H=(1,
               0), D=(1,-1) (the UH's are shown between parentheses).
%e A114576 Triangle begins:
%e A114576 1;
%e A114576 1;
%e A114576 2;
%e A114576 3,1;
%e A114576 6,3;
%e A114576 11,10;
%e A114576 23,26,2;
%e A114576 47,70,10;
%p A114576 G:=(1-z-sqrt(1-2*z-3*z^2-4*z^3*t+4*z^3))/2/z^2/(1-z+t*z): Gser:=simplify(series(G,
               z=0,20)): P[0]:=1: for n from 1 to 16 do P[n]:=coeff(Gser,z^n) od: 
               for n from 0 to 16 do seq(coeff(t*P[n],t^j),j=1..1+floor(n/3)) od; 
               # yields sequence in triangular form
%Y A114576 Cf. A001006, A090344, A014531.
%Y A114576 Sequence in context: A162984 A166295 A016730 this_sequence A116468 A110237 
               A076631
%Y A114576 Adjacent sequences: A114573 A114574 A114575 this_sequence A114577 A114578 
               A114579
%K A114576 nonn,tabf
%O A114576 0,3
%A A114576 Emeric Deutsch (deutsch(AT)duke.poly.edu), Dec 09 2005