3
$\begingroup$

Planning of the question:

Let $(M,g)$ be a Riemannian manifold and $TM$ be its tangent bundle

The isotropic almost complex structures $J_{\delta , \sigma}$ were introduced by Aguilar on the tangent bundle of a Riemannian manifold $(M,g)$ in this paper. These structures induce a class of Riemannian metrics $g_{\delta ,\sigma}$ on $TM$ which are a generalization of the Sasaki metric ($\sigma, \delta :TM \to \mathbb{R}^+$ are real positive functions).

I supposed that the base manifold is the Euclidean space $\mathbb{R}^n$ and the Riemannian manifold $(T\mathbb{R}^n,g_{\delta ,0})$ is an Einstein manifold with the Ricci operator $Ric(A)=\rho A$ where $A$ is a tengent vector to the manifold $T\mathbb{R}^n$ and $\rho =constant$. Then, I concluded the following necessary conditions for $\delta = \frac{1}{\alpha}$:

‎\begin{align}‎ ‎&\sum _{i=1}^n\frac{\partial ^2\alpha}{\partial (y^i)^2 }=-2\rho‎ , ‎\\‎ ‎&\frac{1-n}{2 \alpha}(\frac{\partial \alpha}{\partial y^j})^2-\frac{\partial ^2 \alpha}{\partial (y^j)^2}=2\rho‎ , ‎\hspace{2mm}j=1,...,n,\\‎ ‎&\frac{1-n}{2 \alpha}\frac{\partial \alpha}{\partial y^j}\frac{\partial \alpha}{\partial y^i}-\frac{\partial ^2 \alpha}{\partial y^j\partial y^i}=0,\hspace{2mm}i,j=1,...,n \hspace{2mm}\textit{and} \hspace{2mm}i\neq j‎,\\ &\frac{\partial \alpha}{\partial x^i}=0,\qquad i=1,...,n, ‎\end{align}‎ where we suppose that $\alpha = \frac{1}{\delta}$ and $(x^1,...,x^n)$ and $(x^1,...,x^n,y^1,...,y^n)$ are the standard coordinate systems on $\mathbb{R}^n$ and $T\mathbb{R}^n$, respectively(the summation on the indices is considered only in the first equation).

Question: Is there any function $\alpha: T\mathbb{R}^n \to \mathbb{R}^+$ satisfying the mentioned system? The locally answers are acceptable, too.

$\endgroup$

1 Answer 1

3
$\begingroup$

Unless I am somehow misunderstanding the notation, we can forget about the x dependence due to your last condition. So let $\alpha$ be a function of y. Your third equation can be written as $${\partial^2 \over\partial y_i\partial y_j}(\alpha^{(n+1)/2})=0.$$ So it follows that $$\alpha^{(n+1)/2}=\sum_i f_i(y_i).$$ Now your second condition says that $${\partial ^2\over \partial y_i^2}(\alpha^{(n+1)/2})=-{4\rho\over (n+1)}\alpha^{(1-n)/2}.$$ Since the left hand side depends only on $y_i$, this is possible only if $\alpha$ is also a function of only $y_i$. But if this applies for every $i$, then $\alpha$ must be constant.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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