9G - Estimating small changes
How can a derivative estimate a nearby square root or the effect of a tiny change in radius? You will use a tangent to estimate changes and new values, then compare estimates with exact calculations so that their accuracy is stated meaningfully.
A tangent as a local approximation
A tangent has the same gradient as a smooth curve at its point of contact, making it a useful approximation very close to that point.
Suppose , and changes from to . The exact change is . On the tangent at , a horizontal change produces a vertical change equal to gradient times .
Small changes
For a sufficiently small change near a differentiable point ,
The specification also writes this as for small . Evaluate the derivative at the starting value. The symbol matters: a curved graph generally leaves an error.
[DIAGRAM: asset_name: A tangent estimates a nearby change; asset_slug: tangent-change; caption: Near x equals 3, the tangent to y equals x squared follows the curve closely; description: The curve y equals x squared and its tangent y equals 9 plus 6 times (x minus 3) are plotted close to (3,9). At x equals 3.2 a vertical gap separates tangent value 10.2 and curve value 10.24. A horizontal arrow marks delta x equals 0.2 and a vertical arrow marks tangent change 1.2. The gap 0.04 is labelled as the omitted square of delta x.]

Keep the sign of . If the input decreases, ; whether the output increases or decreases then depends on the gradient as well.
Worked example 1: a small increase
Question. For , estimate the change in when increases from to .
Plan. Use the gradient at the starting point, , and the small input change .
Hence
Answer. The estimated increase in is .
Check. The exact difference can be factored: . The estimate is close because the secant gradient is close to the starting tangent gradient .
Now reverse the input change while keeping its size and the starting point. The derivative is unchanged, but the estimated change must reverse sign.
Worked example 2: a small decrease
Question. For , estimate the change in when decreases from to .
Plan. Use the same starting gradient , now with .
Answer. The estimated change is , a decrease of .
Check. Exactly, . Its sign agrees with the estimate. The secant gradient is again close to .
A change is a signed difference, not the new value of the function.
Estimating a change or a new value
An estimated change must be added to the starting quantity if the question asks for its new value.
Choose a nearby starting value where the function and derivative are easy to evaluate. For a square root, a nearby square number is useful. For percentage changes, first turn the percentage input change into an actual change, or divide the small-change equation by the starting output.
Worked example 3: a nearby square root
Question. Use differentiation at to estimate .
Plan. Take , estimate its change from to , then add that change to .
Therefore
Answer. .
Check. , very close to the required . The exact root is therefore slightly below . The estimate is a new root value, whereas is only its estimated increase.
For a percentage increase in a geometric dimension, the output percentage need not be the same. A volume depends on three powers of a length.
Worked example 4: a percentage radius change
Question. A sphere initially has radius . Its radius increases by . Use differentiation to estimate the increase in volume and the percentage increase in volume. Use .
Plan. Convert the radius percentage to , multiply by at , and compare with the original volume.
Thus
The starting volume is , so
Answer. Increase approximately , or of the original volume.
Check. The exact fractional increase is , or . This is close to . Algebraically, dividing by gives , explaining the factor of three.
Continue this lesson.
You’re previewing NovaLearn without an account. Create a free account to read the rest.
Create free account