Lagrange Inversion
Formal Laurent Series
We already know the ring of formal power series \(\mathbb{C}\lbrack\lbrack x\rbrack\rbrack\). Now define the ring of formal Laurent series:
We can define the multiplicative inverse for elements in \(\mathbb{C}\left(\left(x\right)\right)\) by imitating the definition for formal power series:
If for \(f:=\sum_{k\geq N}f_kx^k\) with \(f_N\neq 0\), there exists \(g=\sum_{k\geq -N}g_kx^k\) such that \(fg=1\), then
Similar to formal power series, we also define for nonzero \(f(x)=\sum_{k\geq N}f_kx^k\):
Clearly, for \(g\neq 0\), we have
Formal Residue
The formal residue is the coefficient of the \(x^{-1}\) term in a formal Laurent series. Denote \(\operatorname{res} f:=\lbrack x^{-1}\rbrack f\).
Lemma: For any formal Laurent series \(f\), \(\operatorname{res} f'=0\).
Proof: Consider the definition of the formal derivative \(\left(x^k\right)'=kx^{k-1}\).
Lemma: For any formal Laurent series \(f,g\), \(\operatorname{res}(f'g)=-\operatorname{res}(fg')\).
Proof: Consider the product rule \((fg)'=f'g+fg'\), so \(0=\operatorname{res}((fg)')=\operatorname{res}(f'g)+\operatorname{res}(fg')\).
Lemma: For a formal Laurent series \(f(x)\neq 0\), \(\operatorname{res}(f'/f)=\operatorname{ord}f\).
Proof: Let \(\operatorname{ord}f=k\). Then
Lemma: For a formal Laurent series \(f\) and a formal power series \(g\neq 0\), \(\operatorname{res}(f)\operatorname{ord}(g)=\operatorname{res}(f(g)g')\).
Proof: By linearity, it suffices to prove the case \(f=x^k\) where \(k\in\mathbb{Z}\). If \(k\neq -1\), then
If \(k=-1\), then
Compositional Inverse
Denote \(A(x)\circ B(x):=A(B(x))\).
Proposition: \(f(x):=\sum_{k\geq 1}f_kx^k\) has a compositional inverse \(f^{\langle -1\rangle}(x)\) if and only if \(f(0)=0\neq f'(0)\), and in this case \(f^{\langle -1\rangle}(x)\) is unique. Furthermore: if \(g(x)=\sum_{k\geq 1}g_kx^k\) satisfies \(f(g(x))=x\) or \(g(f(x))=x\), then \(g(x)=f^{\langle -1\rangle}(x)\).
Proof: Consider
Since \(g(f(x))=x\), we have the following system of equations
We can solve the first equation only when \(f_1\neq 0\), then we can solve for \(g_2,\dots\) successively.
In particular, consider \(f(h(x))=x\). Then \(g(f(h(x)))=g(x)\), and thus \(g(x)=g\circ f\circ h(x)=x\circ h(x)=h(x)\).
Lagrange Inversion Formula
Let \(f(x),g(x)\in\mathbb{C}\lbrack\lbrack x\rbrack\rbrack\) satisfy \(f(g(x))=g(f(x))=x\). Take \(\Phi(x)\in\mathbb{C}\lbrack\lbrack x\rbrack\rbrack\) (or \(\Phi(x)\in\mathbb{C}\left(\left(x\right)\right)\)). Then
Proof:
Some readers might be more familiar with the following version: For \(k\in\mathbb{Z}_{\geq 0},n\in\mathbb{Z}_{>0}\),
or
Observe that
This can be derived from the part we have already proven.
References
- Richard P. Stanley and Sergey P. Fomin. Enumerative Combinatorics Volume 2 (Edition 1).
- Ira M. Gessel. Lagrange Inversion.
本页面最近更新:,更新历史
发现错误?想一起完善? 在 GitHub 上编辑此页!
本页面贡献者:OI-wiki
本页面的全部内容在 CC BY-SA 4.0 和 SATA 协议之条款下提供,附加条款亦可能应用