I want to demonstrate:

Finite sum of sigma-finite measures is a sigma finite measure.

I try this: Suppose that are m sigma measures in a space ($\mathcal{X}$,$\mathcal{A}$), denoted by $\mu_k$ with $k \in {1,...,m} $. Then exist $A_n^k$ for each k,and $\mathcal{X}=\bigcup\limits_{n \in \mathbb{N}}A_n^k$.

And i want to prove that $A_n=\bigcup \limits_{k=1}^{m} A_n^k$ is the decomposition of the sum measure $\mu=\sum \limits_{k=1}^{m}\mu_k$, but I have problems with the finite measure of each $A_n$.

My definition of $A_n$ is wrong?

Is an easier solution?



Let us do the case $m=2$ (the general one follows by induction).

The problem is that we may have $\mu_1(A_n^2)=+\infty$.

However, if we define $B_{n_1,n_2}:=A_{n_1}^1\cap A_{n_2}^2$, then the family $\left(B_{n_1,n_2}\right)_{(n_1,n_2)\in\mathbb N^2}$ is countable and $$(\mu_1+\mu_2)\left(B_{n_1,n_2}\right) \leqslant \mu_1\left(A_{n_1}^1\right)+\mu_2\left(A_{n_2}^2\right),$$ which is finite.


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.