Topic 7

Partial Derivatives

fxf_x is the ordinary derivative in xx, with every other variable held fixed as a constant. fyf_y is the same idea in yy. Everything on this page uses that one idea: the limit definition, implicit partials, higher-order partials and .

5 kinds of question. Open the one you are stuck on.

1

Find the first partial derivatives of a function at a point

e.g. “Find all first partial derivatives of the following functions and evaluate them at the given point”
  1. 1
    For fxf_x, treat every other variable as a constant

    Then differentiate in xx with the usual one-variable rules. A factor with no xx in it is just a constant multiple.

  2. 2
    Repeat for each remaining variable
  3. 3
    Plug in the point last

Remember: Freeze the others, differentiate one.

See it in a problem3 worked examples

Let f(x,y)=x3y−4xy2+2yf(x, y) = x^3y - 4xy^2 + 2y. Find fxf_x and fyf_y, and evaluate them at (1,−1)(1, -1).

  1. 1
    For fxf_x, treat every other variable as a constant

    fx=3x2y−4y2f_x = 3x^2y - 4y^2 (the term 2y2y is a constant in xx).

  2. 2
    Repeat for each remaining variable

    fy=x3−8xy+2f_y = x^3 - 8xy + 2.

  3. 3
    Plug in the point last

    fx(1,−1)=−3−4=−7f_x(1, -1) = -3 - 4 = -7;  fy(1,−1)=1+8+2=11\ f_y(1, -1) = 1 + 8 + 2 = 11.

Answer:

fx=3x2y−4y2f_x = 3x^2y - 4y^2, fy=x3−8xy+2f_y = x^3 - 8xy + 2; fx(1,−1)=−7f_x(1, -1) = -7, fy(1,−1)=11f_y(1, -1) = 11

Common mistakes3
  • Differentiating the "constant" too: in xx, y3y^3 is a constant, so ∂∂x(x2y3)=2xy3\dfrac{\partial}{\partial x}(x^2y^3) = 2xy^3, not 2xy3+3x2y22xy^3 + 3x^2y^2.
  • Dropping the chain-rule factor: ∂∂xexy=y exy\dfrac{\partial}{\partial x} e^{xy} = y\,e^{xy}.
  • Plugging in before differentiating: f(1,y)f(1, y) has no xx left to differentiate.
2

Find a partial derivative from the limit definition

e.g. “Compute, directly from the definitions, ∂f∂x(0,0)\dfrac{\partial f}{\partial x}(0, 0)”
  1. 1
    Write fx(a,b)=lim⁡h→0f(a+h,b)−f(a,b)hf_x(a, b) = \displaystyle\lim_{h\to 0}\dfrac{f(a + h, b) - f(a, b)}{h}

    For fyf_y, move only yy: f(a,b+h)−f(a,b)h\dfrac{f(a, b + h) - f(a, b)}{h}.

  2. 2
    Take f(a,b)f(a, b) from the special case

    It is the value given separately for the point itself.

  3. 3
    Use the general formula for f(a+h,b)f(a + h, b)

    For h≠0h \ne 0 the point (a+h,b)(a + h, b) is not the special one.

  4. 4
    Simplify and take the limit

Remember: Move one variable by h, nothing else.

See it in a problem3 worked examples

Let f(x,y)=x3+2y3x2+y2f(x, y) = \dfrac{x^3 + 2y^3}{x^2 + y^2} for (x,y)≠(0,0)(x, y) \ne (0, 0) and f(0,0)=0f(0, 0) = 0. Use the definition to find fx(0,0)f_x(0, 0) and fy(0,0)f_y(0, 0).

  1. 1
    Write fx(a,b)=lim⁡h→0f(a+h,b)−f(a,b)hf_x(a, b) = \displaystyle\lim_{h\to 0}\dfrac{f(a + h, b) - f(a, b)}{h}

    fx(0,0)=lim⁡h→0f(h,0)−f(0,0)hf_x(0, 0) = \displaystyle\lim_{h\to 0}\dfrac{f(h, 0) - f(0, 0)}{h},  fy(0,0)=lim⁡h→0f(0,h)−f(0,0)h\ f_y(0, 0) = \displaystyle\lim_{h\to 0}\dfrac{f(0, h) - f(0, 0)}{h}.

  2. 2
    Take f(a,b)f(a, b) from the special case

    f(0,0)=0f(0, 0) = 0.

  3. 3
    Use the general formula for f(a+h,b)f(a + h, b)

    f(h,0)=h3h2=hf(h, 0) = \dfrac{h^3}{h^2} = h and f(0,h)=2h3h2=2hf(0, h) = \dfrac{2h^3}{h^2} = 2h.

  4. 4
    Simplify and take the limit

    fx(0,0)=lim⁡h→0hh=1f_x(0, 0) = \displaystyle\lim_{h\to 0}\dfrac{h}{h} = 1,  fy(0,0)=lim⁡h→02hh=2\ f_y(0, 0) = \displaystyle\lim_{h\to 0}\dfrac{2h}{h} = 2.

Answer:

fx(0,0)=1f_x(0, 0) = 1,  fy(0,0)=2\ f_y(0, 0) = 2

Common mistakes3
  • Differentiating the formula with the usual rules and plugging in (0,0)(0, 0): the formula does not hold at that point.
  • Moving both variables: for fxf_x, yy stays at bb.
  • Concluding ff is continuous because fxf_x and fyf_y exist. They can exist even where ff jumps.
3

Find partial derivatives implicitly

e.g. “A surface z(x,y)z(x, y) is defined by … Compute ∂z∂x\dfrac{\partial z}{\partial x}, ∂z∂y\dfrac{\partial z}{\partial y} in terms of xx, yy, zz”
  1. 1
    Check the point satisfies the equation

    Then zz is a function of xx and yy near it, as long as Fz≠0F_z \ne 0 there (see step 3).

  2. 2
    Differentiate both sides in xx, holding yy fixed

    Each time you differentiate a zz, multiply by ∂z∂x\dfrac{\partial z}{\partial x}. Product and chain rules still apply.

  3. 3
    Solve for ∂z∂x\dfrac{\partial z}{\partial x}, then repeat for yy

    Shortcut: write the equation as F(x,y,z)=0F(x, y, z) = 0; then ∂z∂x=−FxFz\dfrac{\partial z}{\partial x} = -\dfrac{F_x}{F_z} and ∂z∂y=−FyFz\dfrac{\partial z}{\partial y} = -\dfrac{F_y}{F_z}.

  4. 4
    Plug in the point

Remember: Every z you touch brings a ∂z/∂x.

See it in a problem3 worked examples

The equation x2z+yz3=10x^2z + yz^3 = 10 defines z(x,y)z(x, y) near (1,1,2)(1, 1, 2). Find ∂z∂x\dfrac{\partial z}{\partial x} and ∂z∂y\dfrac{\partial z}{\partial y} there.

  1. 1
    Check the point satisfies the equation

    1⋅2+1⋅8=101 \cdot 2 + 1 \cdot 8 = 10. ✓

  2. 2
    Differentiate both sides in xx, holding yy fixed

    In xx:  2xz+x2zx+3yz2zx=0\ 2xz + x^2 z_x + 3yz^2 z_x = 0.

  3. 3
    Solve for ∂z∂x\dfrac{\partial z}{\partial x}, then repeat for yy

    zx=−2xzx2+3yz2z_x = -\dfrac{2xz}{x^2 + 3yz^2}. In yy: x2zy+z3+3yz2zy=0x^2z_y + z^3 + 3yz^2z_y = 0, so zy=−z3x2+3yz2z_y = -\dfrac{z^3}{x^2 + 3yz^2}.

  4. 4
    Plug in the point

    x2+3yz2=1+12=13x^2 + 3yz^2 = 1 + 12 = 13:  zx=−413\ z_x = -\tfrac{4}{13},  zy=−813\ z_y = -\tfrac{8}{13}.

Answer:

∂z∂x=−413\dfrac{\partial z}{\partial x} = -\dfrac{4}{13},  ∂z∂y=−813\ \dfrac{\partial z}{\partial y} = -\dfrac{8}{13}

Common mistakes3
  • Treating zz as a constant: ∂∂x(z3)=3z2 ∂z∂x\dfrac{\partial}{\partial x}\big(z^3\big) = 3z^2\,\dfrac{\partial z}{\partial x}, not 00.
  • Losing the minus sign in −FxFz-\dfrac{F_x}{F_z}.
  • Using the shortcut without moving everything to one side first.
4

Find second and higher-order partial derivatives

e.g. “Find all second partial derivatives of f(x,y)=…f(x, y) = \dots”
  1. 1
    Read the order of differentiation

    fxyf_{xy}: first xx, then yy (left to right). ∂2f∂y ∂x\dfrac{\partial^2 f}{\partial y\,\partial x}: also xx first (right to left).

  2. 2
    Use Clairaut to pick the easiest order

    If the partials are continuous, gives fxy=fyxf_{xy} = f_{yx}. A third-order mixed partial is then the same in any order.

  3. 3
    Differentiate one variable at a time
  4. 4
    Simplify, then plug in the point if asked

Remember: Subscripts left to right; order is free (Clairaut).

See it in a problem4 worked examples

Find all second partial derivatives of f(x,y)=x2e3yf(x, y) = x^2e^{3y}.

  1. 1
    Read the order of differentiation

    Four of them: fxxf_{xx}, fxyf_{xy}, fyxf_{yx}, fyyf_{yy}.

  2. 2
    Use Clairaut to pick the easiest order

    Clairaut: fyx=fxyf_{yx} = f_{xy}, so only three need computing.

  3. 3
    Differentiate one variable at a time

    fx=2xe3yf_x = 2xe^{3y},  fy=3x2e3y\ f_y = 3x^2e^{3y}; then fxx=2e3yf_{xx} = 2e^{3y},  fxy=6xe3y\ f_{xy} = 6xe^{3y},  fyy=9x2e3y\ f_{yy} = 9x^2e^{3y}.

  4. 4
    Simplify, then plug in the point if asked

    Already simple.

Answer:

fxx=2e3yf_{xx} = 2e^{3y},  fxy=fyx=6xe3y\ f_{xy} = f_{yx} = 6xe^{3y},  fyy=9x2e3y\ f_{yy} = 9x^2e^{3y}

Common mistakes2
  • Reading ∂2f∂y ∂x\dfrac{\partial^2 f}{\partial y\,\partial x} as "yy first". Read the bottom right to left: xx goes first.
  • Assuming ff exists because fxf_x and fyf_y are given: check (fx)y=(fy)x(f_x)_y = (f_y)_x first.
5

Show a function satisfies a partial differential equation

e.g. “Show that uu satisfies the heat equation (wave equation, Laplace's equation)”
  1. 1
    List the partials the PDE needs
  2. 2
    Compute each one
  3. 3
    Substitute into the PDE and simplify
  4. 4
    Conclude, or solve for the unknown constant

    It must hold for every input: cancel a factor that is never 00 (like e−kte^{-kt}), then compare what is left.

Remember: Compute, substitute, compare.

See it in a problem3 worked examples

Show that u(x,y)=xx2+y2u(x, y) = \dfrac{x}{x^2 + y^2} is harmonic: uxx+uyy=0u_{xx} + u_{yy} = 0.

  1. 1
    List the partials the PDE needs

    uxxu_{xx} and uyyu_{yy}.

  2. 2
    Compute each one

    ux=y2−x2(x2+y2)2u_x = \dfrac{y^2 - x^2}{(x^2 + y^2)^2},  uxx=2x3−6xy2(x2+y2)3\ u_{xx} = \dfrac{2x^3 - 6xy^2}{(x^2 + y^2)^3};  uy=−2xy(x2+y2)2\ u_y = -\dfrac{2xy}{(x^2 + y^2)^2},  uyy=6xy2−2x3(x2+y2)3\ u_{yy} = \dfrac{6xy^2 - 2x^3}{(x^2 + y^2)^3}.

  3. 3
    Substitute into the PDE and simplify

    uxx+uyy=2x3−6xy2+6xy2−2x3(x2+y2)3=0u_{xx} + u_{yy} = \dfrac{2x^3 - 6xy^2 + 6xy^2 - 2x^3}{(x^2 + y^2)^3} = 0.

  4. 4
    Conclude, or solve for the unknown constant

    It holds at every (x,y)≠(0,0)(x, y) \ne (0, 0), so uu is harmonic.

Answer:

uxx+uyy=0u_{xx} + u_{yy} = 0 for all (x,y)≠(0,0)(x, y) \ne (0, 0), so uu is harmonic.

Common mistakes3
  • Checking a single point. A PDE has to hold for every input.
  • Canceling a factor that can be 00, like sin⁡(3x)\sin(3x), without saying the equation holds there anyway.
  • Confusing uxxu_{xx} (differentiate twice) with (ux)2(u_x)^2 (square the derivative).

Need the formulas? Must memorize · Flashcards

Found this helpful? Leave us a review on Google ↗Spotted a mistake?

Which topics are you reviewing?

The topic list, Flashcards, Must memorize and Exam words will only show these. Change it any time with the Topics button above the topic list.

Vectors & geometry
Partial derivatives
Optimization
Multiple integrals
11 of 20 topics selected