Education
 

Integration by parts

From Mathematics

In calculus, integration by parts is a technique of integration applicable to integrands consisting of a product that cannot be rewritten as one or more easily integrated terms — at least, not without difficulty. The technique is particularly useful in cases containing a product of algebraic and transcendental factors.

Given two differentiable functions math and math,

math.

To use the technique, one identifies suitable functions math and math and then differentiates math to get math and integrates math to get math — ignoring the usual constant of integration term ("math"), since it does not affect the final answer.

Note that the rule can also be written

math,

for differentiable functions math and math.

Contents

[edit] Derivation

According to the product rule for differentiation, given two differentiable functions math and math,

math (alternatively, math).

Therefore,

math,

for suitably chosen antiderivatives. Simplifying the right-hand side of the equation,

math.

On the left-hand side, the integral clearly "undoes" the differentiation (by the fundamental theorem of calculus), so

math.

Because the two antiderivative terms can always be chosen to make math, this can be simplified to:

math.

Solving for math, one obtains the final form of the rule:

math.

[edit] Examples

[edit] Polynomial factor to a large power

A fairly simple example of integration by parts is the integral

math.

Although the integrand only involves algebraic functions, it is a good candidate for the method because expansion of math would be very tedious.

The key to the successful use of integration by parts is finding a usable value for math. Doing so is something of an art and may require trial and error.

First consider a wrong way to do this integral by parts:

  • Let math and math (since it is the term left over after math is determined).
  • Thus math and math.

Then

math.

However, it would be difficult to integrate the second term of the right-hand side of the equation, so this approach will be abandoned.

Here is a better way to handle this case:

  • Let math and math.
  • Thus math and math.

Then

math
math.

Here integration by parts works quite nicely. It can be easily confirmed by differentiation that the resulting antiderivative is correct.

Note that this integral may also be evaluated using the simpler integration by substitution technique.

[edit] Algebraic and transcendental factors

As another example where integration by parts is useful (and, in fact, necessary), consider the integral

math.

Choosing math fails, as in the previous (counter)example, since the resulting integral is more difficult than the original. Instead:

  • Let math and math.
  • Thus math and math.

Then

math
math
math.

In this case, the second term in the final expression requires another application of integration by parts:

  • Let math and math.
  • Thus math and math.

Then

math
math
math.

Substituting the last expression into the previous result:

math.

Note that if the second integration by parts step had instead used math and math, this would have "undone" the first step and we would have ended up with an integrand very much like the one we started with:

math
math
math.

Thus, the correct choice of math and math is particularly important when multiple applications of the technique are required. In general, if math is chosen to be an algebraic function in the first step, it should be algebraic in all subsequent steps.

[edit] Choosing u and dv

Fortunately, there is a mnemonic for choosing math and math, which covers a large variety of integrands:

mathL I A T Emath

The letters stand for:

This mnemonic only works when the integrand is the product of two different types of factors. The factor whose type of function appears higher in this list should generally be chosen as math, the factor whose type appears lower as math.

For example, in the integral

math,

the choices should be math, since this is a logarithmic function, and math, since this is an algebraic one ("L" appears before "A" in the mnemonic). On the other hand, in the integral

math

the proper choices are math (algebraic) and math (exponential).

Note that the second example above also follows the rule suggested by this mnemonic.

If the mnemonic doesn't seem to work for a given integral it is possible that it may be a simpler form that can be evaluated using the substitution method, or perhaps rewritten into a simpler form using algebraic or trigonometric techniques (e.g., trigonometric identities).

A slightly different mnemonic that works almost as well — and has the added benefit of sounding more like an English word — is:

mathL I P E Tmath

Here the "P" stands for Power, which includes polynomials and roots (fractional powers). The other letters are as above.

Notice that the last two letters are switched in this form; this is usually not an issue, since integrals involving a product of trigonometric and exponential factors can generally be done "either way" (with respect to the choice of math and math) or not at all using this technique.