Elementary functions

 Below we list some frequently used elementary functions.

  1. Algebraic functions
  2. Exponential functions
  3. Logarithm
  4. Trigonometric functions
  5. Inverse trigonometric functions
  6. Hyperbolic functions


Algebraic functions

A polynomial of x

f(x)=anxn+an1xn1++a1x+a0

where an,an1,,a0R is a continuous function on R. Such functions are called polynomial functions.

If g(x) and h(x) are polynomial functions such that h(x)0, the function f(x) defined by

f(x)=g(x)h(x)

is continuous on {xxR,h(x)0}. Such functions are called rational functions. If h(x)=1 then f(x)=g(x) so any polynomial functions are also rational functions (i.e., polynomial functions are a special case of rational functions).

We can ``algebraically'' define functions other than polynomial or rational functions. For example, f(x)=x is not a rational function, but satisfies an algebraic equality

[f(x)]2x=0.

Definition (Algebraic function)

The continuous function f(x) is said to be an algebraic function if there exist polynomial functions g0(x),g1(x),,gn(x) such that the following identity is satisfied:

gn(x)[f(x)]n+gn1(x)[f(x)]n1++g1(x)f(x)+g0(x)=0.

Example. Rational functions are a special case of algebraic functions. □

Example. Let us prove that the function f(x)=x defined on x0 is continuous.

First, consider the case when a>0. For all ε>0, let us define δ=εa. Then, if x0 and 0<|xa|<δ,

|xa|=|(xa)(x+a)x+a|=|xa|x+a|xa|a<δa=ε.

Next, consider the case when a=0. For any ε>0, take δ=ε2>0. If 0<x<δ, then |x0|=x<δ=ε. Thus, limx+0x=0. Therefore, f(x)=x is continuous at all x0. □

Exponential functions

Definition (Exponential function)

Let a>0 be a real number. The function defined by

f(x)=ax

is called an exponential function with base a. In particular, when we simply say the exponential function, the base is e (Napier's constant).

Exponential functions are continuous everywhere on R. If a>1, then ax is a strictly increasing function. If 0<a<1, then ax is a strictly decreasing function. If a=1, then ax=1 for all xR.

But what do we mean by ax, exactly? Review how we introduced ex through exp(x).

See also: log and e

Logarithm

First, we provide the following theorem without proof.

Theorem

  1. Let f(x) be a continuous and strictly monotone increasing function defined on an interval I. Then f(x) has the inverse f1(x) which is also continuous and strictly monotone increasing.
  2. Let f(x) be a continuous and strictly monotone decreasing function defined on an interval I. Then f(x) has the inverse f1(x) which is also continuous and strictly monotone decreasing.
Remark. If f(x) is a strictly monotone function (either increasing or decreasing) on an interval I, then the function is a bijection from I to f(I). For every bijection, there exists an inverse map that is also bijective. □

Using this theorem, we can see that for each exponential function f(x)=ax, there is its inverse function f1(x) which we define as the logarithmic function with base a denoted f1(x)=loga(x). When the base is e (Napier's constant), loge(x), this is the logarithm we have defined earlier and we often omit the base to write simply log(x) or ln(x).

Theorem 

We have the following limits.
  1. limx0log(1+x)x=1.
  2. limx0ex1x=1.
Proof
  1. limx0log(1+x)x=limx0log(1+x)1x=loge=1.
  2. Let t=ex1. Then x=log(1+t) As x0, t0 so limx0ex1x=limt0tlog(1+t)=1 using the result of Part 1.

Trigonometric functions

We already know sin and cos. The tangent function is defined as
tanx=sinxcosx, xR{(n+12)πnZ}.
Quiz. Why do we exclude the points x=(n+12)π,nZ, from the domain of the tangent function? □

sin and cos have the fundamental period of 2π whereas tan has the fundamental period of π. That is, for nZ,
sin(x+2πn)=sinx,cos(x+2πn)=cosx,tan(x+πn)=tanx.

We give here the following result without proof (for the moment):
limx0sinxx=1.
Example. limx01cosxx2=limx0(1cosx)(1+cosx)x2(1+cosx)=limx01cos2xx2(1+cosx)=limx0(sinxx)211+cosx=12.

Inverse trigonometric functions

Since sin, cos, and tan are periodic functions (and hence not monotone), they don't have inverse functions. Nevertheless, by restricting their domains, we may define the inverse functions.
sinx is strictly monotone increasing on the closed interval [π2,π2]. Therefore it has an inverse function on this domain, which we define as arcsinx. In other words, we consider sinx as a function
sin:[π2,π2][1,1]
so its inverse function is
arcsin:[1,1][π2,π2].

Similarly, we restrict the domain of cosx to [0,π] to define its inverse, which we call arccosx:
cos:[0,π][1,1]
and
arccos:[1,1][0,π].

We restrict the domain of tanx to the open interval (π2,π2) to define its inverse, which we call arctanx:
tan:(π2,π2)(,),
and
arctan:(,)(π2,π2).

Note that these definitions of inverse trigonometric functions are not unique. We could restrict the domains of the trigonometric functions differently. For example, we could restrict the domain of tan as
tan:(π2,3π2)(,)
to define
arctan:(,)(π2,3π2).

Example. Let us find the value of arcsin(sin5π6). First note that sin5π6=sinπ6 (why?) and π6[π2,π2]. Therefore arcsin(sin5π6)=π6. □

Hyperbolic functions

The hyperbolic cosinehyperbolic sine, and hyperbolic tangent are defined, respectively, by
coshx=ex+ex2,sinhx=exex2,tanhx=sinhxcoshx=exexex+ex.
The domain of these functions is R.

Why are their names similar to trigonometric functions? Note that
cosh2θsinh2θ=1.
Recall that the equation x2y2=1 represents the unit hyperbola on R2
See also: Unit hyperbola (Wikipedia) 
Compare this with
cos2θ+sin2θ=1,
and x2+y2=1 represents the unit circle on R2.
Therefore, on R2, while (cosθ,sinθ) corresponds to a point on the unit circle centered at the origin, (coshθ,sinhθ) corresponds to a point on the unit hyperbola.
Also, compare the following relations with the definitions of the hyperbolic functions: 
cosx=eix+eix2,sinx=eixeix2i,tanx=sinxcosx=eixeixi(eix+eix).
Example
1sinh2x1tanh2x=4(exex)2(ex+ex)2(exex)2==1.
(You should fill in the details.) □

Example
sinh(x+y)=sinhxcoshy+coshxsinhy,
and
sinh(xy)=sinhxcoshycoshxsinhy.
(You should prove these equalities.) □

Comments

Popular posts from this blog

Birth process

Branching processes: Mean and variance

Informal introduction to formal logic