If we have a finite measure space and $\{M_{j}\}_{j=1}^{\infty} \in \mathbb{M}$ where $\mathbb{M}$ is a sigma algebra. Then the following holds

$$\mu\Big(\bigcup_{j=1}^{n}M_{j}\Big) = \sum_{\emptyset \neq K \subset \{1,\dots,n\}}(-1)^{|K|-1}\mu\Big(\bigcap_{j \in K}M_{j}\Big)$$

Proposed approach:

It seems that this formula is similar to an inclusion-exclusion formula? One approach I was thinking was an induction approach. Obviously if we take $|K|=1$ the formula holds. The induction step could be to assume it holds for $|K-1|-1$ and then simply prove the final result.

Does this seem a viable approach, any other suggested approaches are welcome. Thanks.


The principle of inclusion/exclusion is actually based on this equality of indicator functions: $$\mathbf{1}_{\bigcup_{i=1}^{n}M_{i}}=\sum_{k=1}^{n}\left(-1\right)^{k-1}\sum_{1\leq i_{1}<\cdots<i_{k}\leq n}\mathbf{1}_{M_{i_{1}}\cap\cdots\cap M_{i_{k}}}\tag1$$

For a proof of that see this answer.

If on both sides an integral is taken w.r.t. to some measure $\mu$ then we arrive at your formula.

| cite | improve this answer | |
  • $\begingroup$ Brilliant. Thank you, good to actually see where the inclusion/exclusion principal is based from. $\endgroup$ – VBACODER Aug 31 '19 at 14:46
  • $\begingroup$ You are very welcome. $\endgroup$ – drhab Aug 31 '19 at 15:12

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.