Measure of Union in Terms of Measure of Intersection

From the definition of a measure, we require \(\mu(A) + \mu(B) = \mu(A \cup B)\) whenever \(A \cap B = \varnothing\). The following result handles cases in which \(A \cap B \neq \varnothing\), that is, they overlap, by introducing an additional term.

Theorem

Given a measure \(\mu\) on the \(\sigma\)-algebra \(\mathcal{F}\), for \(A, B \in \mathcal{F}\):

\[ \mu(A) + \mu(B) = \mu(A \cap B) + \mu(A \cup B).\]

The primary utility of this result is to prove the measure theory inclusion exclusion principal, with the base case there being a more intuitively obvious version of the above

\[ \mu(A \cup B) = \mu(A) + \mu(B) - \mu(A \cap B)\]

and resembling the standard inclusion exclusion principle, where measuring the union of the two sets is equivalent to measuring them separately, and then accounting for double counting the overlap:

Proof

Consider that:

\[ A \cup B = A \cup (B - A) \quad \text{and} \quad A \cup B = B \cup (A - B)\]

and hence:

\[ \mu(A \cup B) = \mu(A) + \mu(B - A) \quad \text{and} \quad \mu(A \cup B) = \mu(B) + \mu(A - B)\]

given that \(A\) and \(B - A\) are disjoint by definition of the set difference, and likewise for \(B\) and \(A - B\).

Now, given the sum of the above expressions in terms of \(\mu(A)\) and \(\mu(B)\) respectively, we have:

\[\begin{align*} \mu(A) + \mu(B) &= \mu(A \cup B) - \mu(B - A) + \mu(A \cup B) - \mu(A - B) \\ &= 2\mu(A \cup B) - (\mu(B - A) + \mu(A - B) + \mu(A \cap B)) + \mu(A \cap B) \\ &= 2\mu(A \cup B) - \mu((B - A) \cup (A - B) \cup (A \cap B) + \mu(A \cap B) \\ &= 2\mu(A \cup B) - \mu(A \cup B) + \mu(A \cap B) \\ &= \mu(A \cup B) + \mu(A \cap B) \\ \end{align*}\]