A304462 Coefficients of the compositionally inverted power series g:=f^{-1} of a formal power series f with the starting coefficients f_0=0 and f_1=1 expressed as polynomials in the coefficients f_2, f_3, ... of the given power series f(X) = X + f_2*X^2 + f_3*X^3 + ...
1, -1, -1, 2, -1, 5, -5, -1, 6, 3, -21, 14, -1, 7, 7, -28, -28, 84, -42, -1, 8, 8, -36, 4, -72, 120, -12, 180, -330, 132, -1, 9, 9, -45, 9, -90, 165, -45, -45, 495, -495, 165, -990, 1287, -429
Offset: 0
Examples
Matrix lexicographically descending in the rows: for instance f(5) f(2)^2 f(1)^3 (-36) > f(4)^2 f(1)^4 (+4) 1; -1; -1,2; -1,5,-5; -1,6,3,-21,14; -1,7,7,-28,-28,84,-42; -1,8,8,-36,4,-72,120,-12,180,-330,132; -1,9,9,-45,9,-90,165,-45,-45,495,-495,165,-990,1287,-429; -1,10,10,-55,10,-110,220,5,-110,-55,660,-715,-55,330,660,-2860,2002,55,-1430,5005,-5005,1430;
References
- Morse, P. M. and Feshbach, H., Methods of Theoretical Physics, Part I. New York: McGraw-Hill, 1953.
Crossrefs
Cf. A111785.
Programs
-
MuPAD
alfa:=["a","b","c","d","e","f","g","h","i","j","k"]: byRow := proc(od, // original weighted degree wd, // remaining weighted degree il, // index of last indeterminate jl, // exponent of last indeterminate ni, // remaining number of indeterminates lx) // lexicographic string local j; begin if wd > 1 then for j from min(wd,il) downto 2 do: if j >= il then j:=il: // stay at the latest indeterminate byRow(od,wd-j+1,j,jl+1,ni-1,lx.alfa[j]): else // advance to next indeterminate byRow(od,wd-j+1,j,1 ,ni-1,lx.alfa[j]): end_if: end_for: else // output the monomial dd:=1: d0:="+": dc:=1: for j from length(lx)-1 downto 0 do: d1:=substring(lx,j): if d1 <> d0 then d0:=d1: dc:=1: dd:=-dd: else // the indeterminate changes dc:=dc+1: dd:=-dd*dc: end_if: end_for: nn:=fact(2*od-ni-2)/fact(od): // rising factorial // One row of A304462: coefficients of the lexicographically descending monomials: print(nn/dd): // One row of A304462: coefficients of the lexicographically descending monomials // plus some representation of the monomials themselves: // for j from 1 to ni do: // lx:=lx."a": // end_for: // print(nn/dd,lx): // monomial lx end_if: end_proc: // Output the 8th row: n:=8: byRow(n,n,n,0,n-1,"")
Formula
g(n) := f(1)^(-n) Sum_{j(2), j(3), ...} (-1)^{j(2) + j(3) + ...} ((n-1 + j(2) + j(3) + ...)!)/(n! j(2)! j(3)! ...) ((f(2))/(f(1))^j(2) ((f(3))/(f(1)))^j(3) ...
The sum is to be taken over all combinations of the exponents {j(2), j(3), j(4), ...} with j(2) + 2j(3) + 3j(4) + ... = n-1. See Morse, P. M. and Feshbach, H. pp. 411-413.
Comments