3
$\begingroup$

I have asked the following question in math.stackexchange, but I could not receive the answer. See here.

Suppose $h_{i\overline{j}}$, where $1\leq i, j\leq n$, are functions defined on $\mathbb{C}^n$ such that $h_{i\overline{j}}$ are smooth when viewed as functions defined on $\mathbb{R}^{2n}$ and $H=(h_{i\overline{j}})$ is a Hermitian positive definite matrix. I wonder if it is always possible to find smooth functions $p_1,..., p_n$ defined on $\mathbb{R}^{2n}$ such that $$\frac{\partial p_i}{\partial z_j}+\overline{\frac{\partial p_j}{\partial z_i}}=h_{i\overline{j}}\mbox{ and } \frac{\partial p_i}{\partial \overline{z}_j} -\frac{\partial p_j}{\partial \overline{z}_i}=0.$$ In a related question, I require $p_i$ to be holomorphic, which seems to be too strong.

$\endgroup$
0
$\begingroup$

I find a possible way to solve it, assuming some properties of polynomials which I will state in the end: Write $p_j(z)=a_j+\sqrt{-1}b_j$ for $1\leq j\leq n$. Equating the real and imaginary parts, we see that solving $(1)$ is equivalent to solving $$\tag{2} \frac{\partial a_j}{\partial x_i}+\frac{\partial b_j}{\partial y_i} +\frac{\partial a_i}{\partial x_j}+\frac{\partial b_i}{\partial y_j}=2h_{i\overline{j}},\\ \frac{\partial b_j}{\partial x_i}-\frac{\partial a_j}{\partial y_i} -\frac{\partial b_i}{\partial x_j}+\frac{\partial a_i}{\partial y_j}=0,\\ \frac{\partial a_j}{\partial x_i}+\frac{\partial b_j}{\partial y_i} -\frac{\partial a_i}{\partial x_j}-\frac{\partial b_i}{\partial y_j}=0 $$ where $1\leq i, j\leq n$. Note that $(2)$ is a system of linear PDEs with constant coefficients. Let \begin{equation} P_{ij}^{k}(X_1,X_2,...,X_{2n-1},X_{2n})=\left\{ \begin{array}{ll} X_{2j-1}, & \hbox{ if $k=2i-1$;} \\ X_{2j}, & \hbox{ if $k=2i$;}\\ X_{2i-1}, & \hbox{ if $k=2j-1$;} \\ X_{2i}, & \hbox{ if $k=2j$;}\\ 0, & \hbox{ otherwise.} \end{array} \right. \end{equation} and $$(u_1,u_2,...,u_{2n-1},u_{2n})=(a_1,b_1,...,a_n,b_n).$$ Then the system of equations in the first line of $(2)$ can be written as $$\sum_{1\leq k\leq l\leq n}P_{ij}^{k}\Big(\frac{\partial}{\partial x_1},\frac{\partial}{\partial y_1},...,\frac{\partial}{\partial x_n},\frac{\partial}{\partial y_n}\Big)u_k=0$$ where $1\leq i\leq j\leq n$.

I guess that if $\{Q_{i,j}\}_{1\leq i\leq j\leq n}$ are polynomials in $X_1$, $X_2$,..., $X_{2n-1}$, $X_{2n}$ such that $$ \sum_{1\leq i\leq j\leq n}Q_{i,j}P_{ij}^k=0\mbox{ for all }1\leq k\leq 2n, $$ then we must $Q_{i,j}=0$ for all $1\leq i\leq j\leq n$, which I have posted it as a question here. If this is true, we can then apply Theorem 7.6.13 in the book of Hormander [ An introduction to complex analysis in several variables, Third edition, North-Holland, Amsterdam, 1990] to conclude that the system $(2)$ is solvable.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.