1
$\begingroup$

Say I have an open cover of the interval $(0,1)$. I'll call this open cover $\{V_j\}_{j \in J}$. On each of these open sets there is a continuous, nonzero, complex-valued function $f_j: V_j \to \mathbb{C} \setminus \{0\}$. Is it possible to find a partition of unity $\{p_j\}_{j \in J}$ subordinate to the open cover $\{V_j\}_{j \in J}$ such that the continuous complex-valued function $\sum_j p_j\, f_j $ is never zero?

If it's possible but not easy to see, then I wouldn't mind if someone just provides reference(s)?

$\endgroup$
0
$\begingroup$

It's not always possible. Here's a counterexample. Take $V_1 = (0,\frac23)$, $V_2 = (\frac13,1)$, $f_1 \equiv 1$, and $f_2\equiv -1$. Then, whatever $p_1$ and $p_2$ are, we have $\sum_j p_j f_j = p_1 - p_2$, which is real-valued. Since $p_1-p_2 \equiv p_1\equiv 1$ on $(0,\frac13)$ and $p_1 - p_2 \equiv -p_2 = -1$ on $(\frac23,1)$, by the intermediate value theorem there must be a point $x\in (\frac13,\frac23)$ where $p_1(x)-p_2(x)=0$.

$\endgroup$

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.