Integration by parts
Integration by parts is the integration analogue of the product rule. It is useful when an integrand is a product and differentiating one factor simplifies it.
Start from the product rule:
Rearrange and integrate:
Here is the part we choose to differentiate, and is the part we choose to integrate.
How to choose
A practical heuristic is LIATE: logarithmic, inverse trig, algebraic, trigonometric, exponential. Pick from the earliest available type, because it usually becomes simpler under differentiation. This is a heuristic, not a theorem.
Worked examples
For , choose and . Then and :
For , treat it as . Choose , , so and :
Definite version
This is common in physics when moving derivatives between factors, for example in variational calculations and energy methods. It is also the one-dimensional ancestor of integration-by-parts identities used with the Del operator ∇ in Vector Calculus index.
Common pitfalls
- Forgetting the minus sign in .
- Choosing so that becomes more complicated.
- Dropping the constant for indefinite integrals.
- Mixing up and : includes the differential, while is its antiderivative.
Quick diagnostic
If repeated integration by parts returns a multiple of the original integral, move that term to the left and solve algebraically. This happens for examples like .