2.4 - Reduction formulae
A high power inside an integral need not mean a long expansion. You will derive relations between neighbouring integrals, choose the starting values they need, and use those relations to find exact definite integrals.
Deriving a recurrence
How could you integrate without expanding it into multiple angles? First make an easier integral appear inside the calculation.
For example, write . Integration by parts differentiates the power and integrates the final sine. The identity then produces both the original integral and one with power two less.
Reduction formula
A reduction formula is a recurrence relating an integral with an integer index to integrals with smaller indices. Together with suitable starting values, it evaluates a whole family of integrals.
The limits belong to the definition of the family. They also determine the boundary term, so retain them throughout the derivation.
Worked example 1: reducing a sine power
Question. For integers , let . Show that for .
Plan. Use parts with and , then replace .
Here and . Thus
The boundary term vanishes: , and because . Collecting the two copies of gives the result.
Answer. , for .
Check. At , the identity gives . Hence , independently of parts.
Change only the upper limit. The method survives, but its endpoint contribution changes.
Worked example 2: keeping a nonzero boundary term
Question. Let . Derive a reduction formula valid for .
Plan. Repeat the same parts calculation with the new upper endpoint.
The product at contains factors of . Moving the final term to the left gives
Answer. , for .
Check. Directly integrating gives . Therefore , as required.
A supplied derivative hint can organise the same argument. For example,
Integrate both sides between the stated endpoints to recover the recurrence. This is the product rule read backwards. Integration by parts is supplied in MF19; these family-specific reduction formulae are not.
Watch out
Dropping before substituting the endpoints loses a term in many recurrences. A boundary term is zero only when its evaluated values show that it is.
Use a different endpoint in the same derivation.
Initial integrals
A recurrence is a route down to an integral you can already evaluate. Before using it, follow the index backwards to identify the necessary starting value.
For a step of two, even and odd indices form separate chains. The even chain ends at ; the odd chain ends at . Neither chain needs a negative index, and a recurrence proved only for cannot be substituted into at .
Worked example 3: two starting values
Question. For , find the initial values needed to evaluate and using the step-two recurrence.
Plan. Trace each chain, then integrate its final member directly.
The chains are and . Hence
Answer. needs ; needs .
Check. The first integral is the area of a rectangle of height and width . For the second, substituting gives .
Now keep the integrand but shorten the interval. Both starting values must be recalculated, just as the boundary term was.
Continue this lesson.
You’re previewing NovaLearn without an account. Create a free account to read the rest.
Create free account