# Linear Combination of Integrals/Definite

## Theorem

Let $f$ and $g$ be real functions which are integrable on the closed interval $\closedint a b$.

Let $\lambda$ and $\mu$ be real numbers.

Then:

- $\displaystyle \int_a^b \paren {\lambda \map f t + \mu \map g t} \rd t = \lambda \int_a^b \map f t \rd t + \mu \int_a^b \map g t \rd t$

## Proof 1

Let $F$ and $G$ be primitives of $f$ and $g$ respectively on $\closedint a b$.

By Linear Combination of Derivatives, $H = \lambda F + \mu G$ is a primitive of $\lambda f + \mu g$ on $\closedint a b$.

Hence by the Fundamental Theorem of Calculus:

\(\displaystyle \int_a^b \paren {\lambda \map f t + \mu \map g t} \rd t\) | \(=\) | \(\displaystyle \bigintlimits {\lambda \map F t + \mu \map G t} a b\) | |||||||||||

\(\displaystyle \) | \(=\) | \(\displaystyle \lambda \bigintlimits {\map F t} a b + \mu \bigintlimits {\map G t} a b\) | |||||||||||

\(\displaystyle \) | \(=\) | \(\displaystyle \lambda \int_a^b \map f t \rd t + \mu \int_a^b \map g t \rd t\) |

$\blacksquare$

## Proof 2

It is clear that for step functions $s$ and $t$:

- $\displaystyle \int_a^b \lambda \map s x + \mu \map t x \rd x = \lambda \int_a^b \map s x \rd x + \mu \int_a^b \map t x \rd x$

Under any partition, the lower sums and upper sums of $f$ and $g$ are step functions, so the above formula relates the lower and upper sums of $f$ and $g$ to the lower and upper sums of the linear combinations of $f$ and $g$.

Because this identity is preserved for all possible partitions of $\closedint a b$, it is preserved for the supremum and infimum of all possible lower and upper sums, so the linear combinations of $f$ and $g$ are integrable.

\(\displaystyle \int_a^b \paren {\lambda \map f t + \mu \map g t} \rd t\) | \(=\) | \(\displaystyle \sup \set {\sum_{\nu \mathop = 1}^n \map {m_\nu^{\paren {\lambda f + \mu g} } } {x_\nu - x_{\nu - 1} }: \forall \nu \in \closedint 1 n x_\nu > x_{\nu - 1} }\) | |||||||||||

\(\displaystyle \) | \(=\) | \(\displaystyle \lambda \sup \set {\sum_{\nu \mathop = 1}^n \map {m_\nu^{\paren f} } {x_\nu - x_{\nu - 1} }: \forall \nu \in \closedint 1 n x_\nu > x_{\nu - 1} } + \mu \sup \set {\sum_{\nu \mathop = 1}^n \map {m_\nu^{\paren g} } {x_\nu - x_{\nu - 1} }: \forall \nu \in \closedint 1 n x_\nu > x_{\nu - 1} }\) | |||||||||||

\(\displaystyle \) | \(=\) | \(\displaystyle \lambda \int_a^b \map f x \rd x + \mu \int_a^b \map g x \rd x\) |

$\blacksquare$

## Sources

- 1967: Tom M. Apostol:
*Calculus Volume 1*: $\S 1.4$