4
$\begingroup$

Let $(X,\Sigma)$ be a measurable space and $(\mu_n)$ a sequence of atomic measures defined on this space. Recall that a measure $\mu$ is atomic if for any measurable $A$ of measure $\mu(A)>0$ there is some measurable $E \subset A$ that is an atom, which means that $\mu(E)>0$, and for any measurable $F\in X$, either $\mu(E\cap F)$ or $\mu(E-F)=0$. Consider the measure $\mu=\sum_{n\in \mathbb{N}} \mu_n$. Is it true that $\mu$ must be atomic? This question is raised (but not answered) here.

EDIT: The answer below by @George reflects the question (here) when it was previously required that the measure not take on infinite value. This is not the correct definition, and does not reflect the original question of Roy A. Johnson.

$\endgroup$
3
  • $\begingroup$ Theorem 2.8 in that paper shows that the answer is yes when the sum-measure is semi-finite. $\endgroup$ May 27, 2014 at 20:56
  • $\begingroup$ Yes in the semi-finite case this is straightforward. $\endgroup$
    – timofei
    May 28, 2014 at 10:54
  • 1
    $\begingroup$ This question is still open... I think. $\endgroup$
    – Squirtle
    Feb 20, 2015 at 20:08

2 Answers 2

2
$\begingroup$

The answer is no. Indeed,for $ k \in N$ let $\mu_k$ be a Dirac measure concentrated at $x=0$(we consider measures on the real axis $R$). Then $\sum_{k \in N}\mu_k$ is not atomic measure in the above-mentioned sense

$\endgroup$
2
  • $\begingroup$ I'm sorry, the definition I gave of an atom is not the good one. We do not ask that the measure of the atom be finite. I apologize for this mistake. $\endgroup$
    – timofei
    May 28, 2014 at 10:54
  • $\begingroup$ This answer is good (given the original question posted here), but it does not answer the (correctly phrased) question originally asked by Roy Johnson in "Atomic and Nonatomic Measures" in the link above. $\endgroup$
    – Squirtle
    Feb 20, 2015 at 20:11
0
$\begingroup$

It seems that the answer is positive: If $(\mu_n)_{n \in \mathbb{N}}$ is a sequence of atomic measures, then the sum $\sum_{n \in \mathbb{N}} \mu_n$ is atomic. This has been shown by P. Capek in his paper The atoms of a countable sum of set functions (Mathematica Slovaca 1989, No. 1, p.81-89; link).

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .