Revision notes

Chapter 8 Differentials and Partial Derivatives: Foundation

The key points, rules to remember, tables and diagrams from all 5 lessons, on one page for quick revision. To keep a copy, use your browser's Print and choose "Save as PDF".

Lesson 1

Linear approximation and absolute, relative and percentage errors

Linear approximation of f at x₀: L(x) = f(x₀) + f′(x₀)(x − x₀). Its graph is the tangent line at (x₀, f(x₀)).

— Definition

Estimating a value by linear approximation

  1. Pick f The function whose value you need, e.g. f(x) = √x.
  2. Pick a nearby x₀ Where f(x₀) and f′(x₀) are easy, e.g. 25.
  3. Find Δx Δx = x − x₀, small.
  4. Compute f(x₀) + f′(x₀) Δx.
MeasureFormulaUnits
Absolute error|actual value − approximate value|same units as the quantity
Relative error|actual − approximate| / |actual| (actual ≠ 0)no units
Percentage errorrelative error × 100per cent

Key takeaways: near x₀, f(x) ≈ f(x₀) + f′(x₀)(x − x₀), the tangent line; pick x₀ so that f(x₀) and f′(x₀) are easy; Δy ≈ f′(x) Δx estimates a change; absolute error has units, relative and percentage errors do not and are the right way to judge an approximation.

Read the full lesson

Lesson 2

Differentials of a function of one variable

Two kinds of change when x moves to x + dx

  • Δy (actual change)
    • Measured along the curve
    • Δy = f(x + dx) − f(x)
    • Exact, often hard to compute
  • dy (differential)
    • Measured along the tangent line
    • dy = f′(x) dx
    • Easy; dy ≈ Δy when dx is small
RuleDifferential form
Constant cdc = 0
Constant multipled(cf) = c df
Sumd(f + g) = df + dg
Productd(fg) = f dg + g df
Quotientd(f/g) = (g df − f dg)/g², g ≠ 0
Composited(f(g(x))) = f′(g(x)) g′(x) dx
Exponentiald(e^f) = e^f f′(x) dx
Logarithm (f > 0)d(log f) = [f′(x)/f(x)] dx

dy = f′(x) dx, with dx = Δx. For small dx, Δy ≈ dy and f(x + dx) ≈ f(x) + f′(x) dx.

— Differentials in one line

Key takeaways: the differential df = f′(x) dx depends on x and on dx; dy is the change along the tangent and Δy the change along the curve; differentials obey the same rules as derivatives; use radians for trigonometric approximations; dividing a differential by the quantity gives the relative change.

Read the full lesson

Lesson 3

Functions of several variables: limits and continuity

Deciding whether lim (x, y) → (u, v) of F exists

  • Direct substitution works F is built from continuous pieces with no zero denominator at (u, v).
    • The limit is F(u, v)
  • Two paths give different values Try lines y = mx, then curves such as x = y².
    • The limit does not exist
  • F − L can be bounded Show |F(x, y) − L| ≤ something that tends to 0.
    • The limit is L

F is continuous at (u, v) if (1) F(u, v) is defined, (2) L = lim ((x, y) → (u, v)) F(x, y) exists, and (3) L = F(u, v).

— Continuity of a function of two variables

Key takeaways: z = F(x, y) is a surface, and fixing one variable gives a curve; neighbourhoods in the plane are open discs; a two-variable limit must be the same along every path, so two paths with different values prove that it does not exist, while a bound |F − L| ≤ (something → 0) proves that it does; continuity needs the value, the limit and their equality.

Read the full lesson

Lesson 4

Partial derivatives, Clairaut's theorem and Laplace's equation

∂F/∂x (x₀, y₀) = lim (h → 0) [F(x₀ + h, y₀) − F(x₀, y₀)]/h; ∂F/∂y (x₀, y₀) = lim (k → 0) [F(x₀, y₀ + k) − F(x₀, y₀)]/k.

— Partial derivatives
SymbolMeaningTextbook subscript
∂²F/∂x²∂/∂x (∂F/∂x)F_xx
∂²F/∂y²∂/∂y (∂F/∂y)F_yy
∂²F/∂x∂y∂/∂x (∂F/∂y): first y, then xF_xy
∂²F/∂y∂x∂/∂y (∂F/∂x): first x, then yF_yx

Checking a function against Laplace's equation

  1. Find u_x and u_y Hold the other variable constant.
  2. Differentiate again u_xx from u_x, u_yy from u_y.
  3. Add u_xx + u_yy.
  4. Decide Zero everywhere: harmonic. Otherwise: not.

Key takeaways: a partial derivative differentiates with respect to one variable while all others are held constant; it is the slope of a section of the surface; partial derivatives can exist without continuity; there are four second order partials, and Clairaut's theorem makes the two mixed ones equal when they are continuous; u is harmonic when u_xx + u_yy = 0.

Read the full lesson

Lesson 5

Linear approximation of several variables, chain rules, homogeneous functions and Euler's theorem

Two variables: L(x, y) = F(x₀, y₀) + F_x(x₀, y₀)(x − x₀) + F_y(x₀, y₀)(y − y₀); dF = F_x dx + F_y dy, with dx = Δx, dy = Δy. Three variables: add F_z(x₀, y₀, z₀)(z − z₀) to L and F_z dz to dF.

— Linear approximation and differential

Tree diagram for W = W(x, y) with x = x(t), y = y(t)

  • Branch through x Multiply ∂W/∂x by dx/dt.
    • W → x → t
  • Branch through y Multiply ∂W/∂y by dy/dt.
    • W → y → t
  • Add the branches dW/dt = W_x x′(t) + W_y y′(t).
FunctionF(λx, λy)Conclusion
x³ + 4x²y − y³λ³(x³ + 4x²y − y³)homogeneous, degree 3
(x² + y²)/(x + y)λ²(x² + y²)/[λ(x + y)]homogeneous, degree 1
√x + √yλ^(1/2)(√x + √y), λ > 0homogeneous, degree 1/2
sin(y/x)sin(λy/λx) = sin(y/x)homogeneous, degree 0
x² + yλ²x² + λynot homogeneous

If F(x, y) is homogeneous of degree p with continuous partial derivatives, then x ∂F/∂x + y ∂F/∂y = pF. For three variables: x F_x + y F_y + z F_z = pF.

— Euler's theorem on homogeneous functions

Key takeaways: L(x, y) = F + F_x(x − x₀) + F_y(y − y₀) is the tangent plane; dF = F_x dx + F_y dy (+ F_z dz); for products, relative errors add; chain rule: add over every path in the tree diagram; F is homogeneous of degree p if F(λx, λy) = λᵖF; Euler: xF_x + yF_y = pF, applied to a homogeneous function of u when u is not homogeneous.

Read the full lesson

Privacy · Terms · Refunds · Delivery