3.6. Step Functions as Integrators#
Suppose \(\alpha\) is constant on \([a, b]\) except possibly at point \(x=a\), that is, \(\alpha(x) = \alpha(b)\) for all \(a < x \leq b\). If \(f\) is continuous from the right at \(a\), then \(f \in \mathfrak{R}(\alpha)\) on \([a, b]\), and
Note
Note that we assume \(\alpha\) possibly has a different value at \(a\). If \(\alpha\) is constant on the entire interval \([a, b]\), the integral clearly exists and is zero.
An analogous result holds when we assume \(\alpha\) is constant on \([a, b]\) except possibly at the right endpoint \(x=b\).
Proof. If \(\alpha(a) = \alpha(b)\), the conclusion is trivial. In the following proof, we assume \(\alpha(a) \neq \alpha(b)\).
Let \(\varepsilon > 0\) be arbitrary. Because \(f\) is continuous at \(x=a\), there exists \(\delta > 0\) such that
Let \(P_\varepsilon = \{x_0, \ldots, x_n\}\) be a partition of \([a, b]\) such that \(x_1 < x_0 + \delta\). For any refinement \(P \supseteq P_\varepsilon\), we have
It then follows that
This completes the proof.
Given \(a < c <b\). Let \(\alpha\) be constant on \([a, b]\) except at point \(x=c\). That is, let \(\alpha(a)\), \(\alpha(c)\) and \(\alpha(b)\) be arbitrary. Define
If function \(f\) is defined in such a way that
At least one of \(f\) and \(\alpha\) is continuous from the left at \(c\), and
at least one of \(f\) and \(\alpha\) is continuous from the right at \(c\),
then \(f \in \mathfrak{R}(\alpha)\) on \([a, b]\), and
Proof. It follows from Proposition 3.2 and its analogous result that \(f \in \mathfrak{R}(\alpha)\) on \([a, c]\) and \(f \in \mathfrak{R}\) on \([c, b]\). The values of the integrals are
Then Theorem 3.3 implies that \(f \in \mathfrak{R}(\alpha)\) on \([a, b]\), and