Notes@HKU by Jax

Derivatives

First principle

The first principle is the definition of the derivative of a function f(x)f(x) at x=ax=a:

f′(x)=ddxf(x)=lim⁡h→0f(x+h)−f(x)hf'(x) = \frac{d}{dx} f(x)=\lim_{h\to 0}\frac{f(x+h)-f(x)}{h}

Note the notation of prime (′'), which denotes the derivative against xx for any function primed.

Differentiability

The function is differentiable at x=ax=a if:

lim⁡x→a−f′=lim⁡x→a+f′\lim_{x\to a^-}f'=\lim_{x\to a^+}f'

Use first principle if the function is cased.

Differentiation formulas and rules

Basic formulas

  • We can differentiate individual items: (f±g)′=f′±g′\displaystyle {\left( {f \pm g} \right)^\prime } = f' \pm g'
  • We can factor out a multiplicative constant: (cf)′=cf′\displaystyle {\left( {cf} \right)^\prime } = cf'
  • Derivative of a constant is 0: ddxk=0\frac{d}{dx}k = 0
  • Power rule: ddxxn=nxn−1\frac{d}{dx}x^n = nx^{n-1}

Chain rule

Shorthand: d1x2 + d2x1

(u(v))′=u′(v)v′ordydx=dydududx(u(v))'=u'(v)v'\quad\quad\text{or}\quad\quad\frac{dy}{dx}=\frac{dy}{du}\frac{du}{dx}

Product rule

Shorthand: d from outside to inside

(uv)′=uv′+vu′orddx(uv)=udvdx+vdudx(uv)' = uv' + vu'\quad\quad\text{or}\quad\quad\frac{d}{dx}(uv)=u\frac{dv}{dx}+v\frac{du}{dx}

Quotient rule

Shorthand: move lower d upper - d lower x upper, lower square

(uv)′=vu′−uv′v2orddx(uv)=vdudx−udvdxv2(\frac{u}{v})'=\frac{vu'-uv'}{v^2}\quad\quad\text{or}\quad\quad\frac{d}{dx}(\frac{u}{v})=\frac{v\frac{du}{dx}-u\frac{dv}{dx}}{v^2}
f(x)f(x)f′(x)f'(x)
1.axa^xln⁡a⋅ax\ln{a}\cdot a^x
2.ekxe^{kx}kekxke^{kx}
3.ln⁡kx\ln{kx}x−1x^{-1}
4.∥x∥\|x\|∥x∥x\frac{\|x\|}{x}
5.sin⁡kx\sin kxkcos⁡kxk\cos kx
6.cos⁡kx\cos kx−ksin⁡kx-k\sin kx
7.tan⁡kx\tan kxksec⁡2kxk\sec^2 kx
8.csc⁡x\csc x−csc⁡xcot⁡x-\csc x \cot x
9.sec⁡x\sec xsec⁡xtan⁡x\sec x \tan x
10.cot⁡x\cot x−csc⁡2x-\csc^2x
11.sin⁡−1x\sin^{-1} x(1−x2)−1(\sqrt{1-x^2})^{-1}
12.cos⁡−1x\cos^{-1} x−(1−x2)−1-(\sqrt{1-x^2})^{-1}
13.tan⁡−1x\tan^{-1} x(1+x2)−1(1+x^2)^{-1}

Techniques of differentiation

Parametric differentiation

For a parametric equation y=f(t)y=f(t) and x=g(t)x=g(t):

dydx=dydt÷dxdt\frac{dy}{dx}=\frac{dy}{dt}\div \frac{dx}{dt}

Inverse functions

f−1(x):f(x)↔xf^{-1}(x): f(x)\leftrightarrow x (switch xx and yy).

  1. Inverse functions only exists in the domain of ff where ff is bijective (one-to-one and onto). This means that the domain where ff gives unique values is the range of f−1f^{-1}.
  2. The range of ff is the domain of f−1f^{-1}.
  3. The domain of ff is the range of f−1f^{-1}.

Derivative of Inverse functions

At cc, we first find f′(x)f'(x) and the value of f−1(c)f^{-1}(c), then we apply the formula to find the value of f−1′(c)f^{-1'}(c):

f−1′(c)=1f′(f−1(c))f^{-1'}(c)=\frac{1}{f'(f^{-1}(c))}

f−1(c)f^{-1}(c) can be found by solving f(x)=cf(x)=c.

Implicit differentiation

Differentiate all xyxy, add y′y' behind all differentiations of yy.

To find y′y' for y2=x2+sin⁡(xy)y^2=x^2+\sin(xy):

y2=x2+sin⁡(xy)2yy′=ddx(sin⁡(xy))2yy′=2x+cos⁡(xy)(xy′+y)Then we simply collect terms of y′y′=2x+ycos⁡(xy)2y−xcos⁡(xy)\begin{aligned} y^2 & =x^2+\sin(xy) \\ 2yy' & =\frac{d}{dx}(\sin(xy)) \\ 2yy' & =2x+\cos(xy)(xy'+y) \\ \text{Then we simply collect terms of } & y' \\ y' & =\frac{2x+y\cos(xy)}{2y-x\cos(xy)} \end{aligned}

To find the second derivative y′′y'', differentate the expression and substitute y′y' back in.

Logarithmic differentiation

To differentiate a function to the power of another function, we can take the logarithm of the function, then differentiate.

(fg)′=(egln⁡f)=(gln⁡f)′egln⁡f\begin{aligned} (f^g)' & =(e^{g\ln f}) \\ & = (g\ln f)'e^{g\ln f} \end{aligned}

To find dydx\frac{dy}{dx} for y=xxy=x^x:

y=exln⁡x=exln⁡x(ln⁡x+x1x)=xx(ln⁡x+1)\begin{aligned} y & = e^{x\ln x} \\ & = e^{x\ln x}(\ln x+x\frac{1}{x}) \\ & = x^x(\ln x+1) \end{aligned}

L'Hopital's rule

For any a∈[R,±∞]a\in [\mathbb{R}, \pm\infty], if lim⁡x→a(fg)\lim_{x\to a}(\frac{f}{g}) is in indeterminate form after substitution, we can conclude:

lim⁡x→a(fg)=lim⁡x→a(f′g′)\lim_{x\to a}(\frac{f}{g})=\lim_{x\to a}(\frac{f'}{g'})

To find lim⁡x→−∞xex\lim_{x\to-\infty}xe^x, we first check if the limit is indeterminate, then we can apply the rule:

lim⁡x→−∞xex  ⟹  ∞×0lim⁡x→−∞xex=lim⁡x→−∞xe−x=lim⁡x→−∞1−e−x (rule applied)=0\begin{aligned} \lim_{x\to-\infty}xe^x & \implies \infty\times0 \\ \lim_{x\to-\infty}xe^x & = \lim_{x\to-\infty}\frac{x}{e^{-x}} \\ & = \lim_{x\to-\infty}\frac{1}{-e^{-x}}\text{ (rule applied)} \\ & = 0 \end{aligned}

Important theorems

Mean value theorem (MVT)

For f(x)f(x) that is continuous in [a,b][a,b] and differentiable in (a,b)(a,b):

f′(c)=f(b)−f(a)b−a,c∈(a,b)f'(c)=\frac{f(b)-f(a)}{b-a},\quad c\in(a,b)

The theorem tells us that, in described conditions, there must be a point cc where the slope of the tangent line is equal to the slope of the line from a→ba\to b (secant line).

Rolle's theorem

For f(x)f(x) that is continuous in [a,b][a,b] and differentiable in (a,b)(a,b):

∀ f(a)=f(b) there exists f′(c)=0,c∈(a,b)\forall\ f(a)=f(b)\text{ there exists }f'(c)=0,\quad c\in(a,b)

For F′=fF'=f, if FF has 4 roots, then ff has 3 roots.

Extremum points

Critical and inflection points

A critical point is a point where f′(x)=0f'(x)=0 or undefined, or the end-points of the domain if inclusive.

An inflection point is a point where f′′(x)=0f''(x)=0 or undefined, that the concavity of the function changes.

Max/minimum points

The absolute maximum/minimum points are the points where the function has the largest/smallest value in the entire domain.

The local maximum/minimum points are the points where the function has the largest/smallest value in a small interval around the point.

Determining shape of graph

Concavity

Concavity is the direction of the curve, and can be described by the values of f′f' and f′′f'':

f′′f''−-++++−-
f′f'−-++−-++
ff

Note: Arrows goes clock-wise.

StepExpression
1.Determine domain of function
2.Special points without continuity?
3.Axis intercepts(f(x)=0,0) (0,f(0))(f(x)=0, 0)\ (0, f(0))
4.Critical pointsf′(x)=0f'(x)=0 or undefined
5.Point maxima?+f′′(x)+f''(x) or −f′′(x)-f''(x)
6.Inflection pointsf′′(x)=0f''(x)=0 or undefined
7.Horizontal asymptoteslim⁡x→±∞f(x)=n?\lim_{x\to\pm\infty}f(x)=n?
8.Vertical asymptoteslim⁡x→a±f(x)=±∞?\lim_{x\to a^\pm}f(x)=\pm\infty?
9.Area strictly increasing/decreasing?f′(x)>0f'(x)>0 or f′(x)<0f'(x)<0
10.Area concavity?−++−/−+−+/-++-/-+-+/

Related: Definition of asymptotes

Higher derivatives

Derivatives of higher order (e.g. f′′(x),f′′′(x)f''(x), f'''(x)) can be expressed as f(n)(x)f^{(n)}(x)

On this page