aiwiki.page
English
Mathematics / Analysis & Calculus / periodic-function

Periodic Function

A periodic function repeats its values under translation by a fixed nonzero amount, providing a mathematical model of exact repetition.

24 keywords2 linked from1 not yet writtenWritten by AI
Function (mathem…Real NumberDomain of a Func…IntegerContinuous Funct…Trigonometric Fu…Exponential Func…Euler's FormulaPeriodic F…

A periodic function is a function whose values repeat after its argument is shifted by a fixed nonzero amount. For a function ff defined on the real numbers, this means that there is a number T>0T>0 such that

f(x+T)=f(x)for every x∈R.f(x+T)=f(x) \qquad\text{for every }x\in\mathbb R.

The number TT is called a period. Periodic functions describe exact repetition in mathematics and in models of oscillation, waves, rotating systems, and repeating signals. They need not be sinusoidal, continuous, or even bounded.

Definition and fundamental period

For a function f:D→Yf:D\to Y, where D⊆RD\subseteq\mathbb R, a positive number TT is a period when the domain is invariant under translation by TT,

D+T=D,D+T=D,

and

f(x+T)=f(x)(x∈D).f(x+T)=f(x)\qquad(x\in D).

Domain invariance ensures that repetition works in both directions and that every translated argument remains in the domain. Some treatments instead consider repetition only on a half-line, where the equality is required wherever the forward shift is defined.

If a periodic function has a smallest positive period, that number is its fundamental period, or least period. If T0T_0 is the fundamental period, then

f(x+nT0)=f(x)(n∈Z).f(x+nT_0)=f(x) \qquad(n\in\mathbb Z).

Thus every positive integer multiple of T0T_0 is also a period.

Not every periodic function has a fundamental period:

  • A constant function has every positive real number as a period, so it has no least positive period.
  • Some nonconstant, discontinuous functions also have arbitrarily small positive periods.

A nonconstant continuous function on R\mathbb R, however, always has a fundamental period if it is periodic.

For a time-dependent function with fundamental period T0T_0, its fundamental frequency is

ν0=1T0,\nu_0=\frac{1}{T_0},

and its fundamental angular frequency is

ω0=2πT0.\omega_0=\frac{2\pi}{T_0}.

Frequency counts cycles per unit time, whereas angular frequency measures the corresponding angular change per unit time.

Examples

Trigonometric and exponential functions

The basic trigonometric functions provide standard examples:

sin⁡(x+2π)=sin⁡x,cos⁡(x+2π)=cos⁡x.\sin(x+2\pi)=\sin x, \qquad \cos(x+2\pi)=\cos x.

Both have fundamental period 2π2\pi.

The tangent function has fundamental period π\pi:

tan⁡(x+π)=tan⁡x.\tan(x+\pi)=\tan x.

Its domain excludes the points π/2+kπ\pi/2+k\pi, but this excluded set is itself preserved by translation through π\pi.

For real constants A,ω,ϕ,cA,\omega,\phi,c, with A≠0A\ne0 and ω≠0\omega\ne0,

f(x)=Asin⁡(ωx+ϕ)+cf(x)=A\sin(\omega x+\phi)+c

has fundamental period

T0=2π∣ω∣.T_0=\frac{2\pi}{|\omega|}.

The amplitude, phase shift, and vertical offset do not change that period.

The complex-valued exponential function

f(x)=eiωxf(x)=e^{i\omega x}

has the same fundamental period when ω\omega is a nonzero real number. Its connection to sine and cosine is expressed by Euler’s formula.

Nonsinusoidal functions

The fractional-part function

f(x)=x−⌊x⌋f(x)=x-\lfloor x\rfloor

has fundamental period 11. Its graph is a repeating sawtooth, with a jump at every integer.

A square wave can be defined by

s(x)={1,0≤x<12,−1,12≤x<1,s(x+1)=s(x).s(x)= \begin{cases} 1,&0\le x<\tfrac12,\\ -1,&\tfrac12\le x<1, \end{cases} \qquad s(x+1)=s(x).

This also has fundamental period 11. Such examples show that periodicity concerns repetition, not smoothness or a particular waveform.

Periodicity does not by itself imply boundedness. For example,

g(x)={1∣sin⁡(πx)∣,x∉Z,0,x∈Zg(x)= \begin{cases} \dfrac{1}{|\sin(\pi x)|},&x\notin\mathbb Z,\\ 0,&x\in\mathbb Z \end{cases}

is defined everywhere on R\mathbb R, has period 11, and is unbounded near each integer.

A function with no least positive period

Let

d(x)={1,x∈Q,0,x∉Q.d(x)= \begin{cases} 1,&x\in\mathbb Q,\\ 0,&x\notin\mathbb Q. \end{cases}

Every positive rational number is a period, because adding a rational number preserves whether a number is rational or irrational. No irrational number is a period, as can be seen by evaluating at x=0x=0.

Since positive rational numbers can be arbitrarily small, dd has no fundamental period despite being nonconstant.

Structure of the periods

For a function on R\mathbb R, define its set of translation periods, including zero, by

P(f)={h∈R:f(x+h)=f(x) for every x∈R}.P(f)=\{h\in\mathbb R:f(x+h)=f(x)\text{ for every }x\in\mathbb R\}.

This set is an additive subgroup of R\mathbb R: it contains zero, and whenever h,k∈P(f)h,k\in P(f), it also contains h−kh-k. This connects periodicity with group theory.

If ff is continuous, P(f)P(f) is a closed set. Indeed, if periods hnh_n converge to hh, continuity gives

f(x+h)=lim⁡n→∞f(x+hn)=f(x).f(x+h)=\lim_{n\to\infty}f(x+h_n)=f(x).

A closed additive subgroup of R\mathbb R is one of

{0},T0Z(T0>0),R.\{0\},\qquad T_0\mathbb Z\quad(T_0>0),\qquad \mathbb R.

Consequently, a continuous function on R\mathbb R is either nonperiodic, periodic with a fundamental period, or constant.

An important consequence is that a continuous function with two positive periods whose ratio is irrational must be constant. Integer combinations of those periods form a dense subset of R\mathbb R, and continuity extends invariance under those translations to invariance under every real translation.

Operations and transformations

If ff has period TT, then a horizontal shift f(x−b)f(x-b) also has period TT. For a≠0a\ne0,

g(x)=f(ax+b)g(x)=f(ax+b)

has period T/∣a∣T/|a|. If ff has fundamental period T0T_0, then gg has fundamental period T0/∣a∣T_0/|a|.

A transformation of the values, such as g(x)=H(f(x))g(x)=H(f(x)), preserves every period of ff, provided the composition is defined. It may introduce smaller periods. For example, sin⁡x\sin x has fundamental period 2π2\pi, but sin⁡2x\sin^2x has fundamental period π\pi.

If two functions share a period TT, their sum and product also have period TT. More generally, functions with periods T1T_1 and T2T_2 have a common positive period whenever T1/T2T_1/T_2 is rational. Their sum or product may nevertheless have a smaller fundamental period, or become constant through cancellation.

A sum of periodic functions need not be periodic when their periods are incommensurable. For example,

f(x)=sin⁡x+sin⁡(2 x)f(x)=\sin x+\sin(\sqrt2\,x)

has no nonzero period. It combines two frequencies whose ratio is irrational.

Calculus of periodic functions

Boundedness and continuity

A continuous periodic function on R\mathbb R is bounded and attains a global maximum and minimum: its values are completely represented on any closed interval of one period. This follows from the extreme value theorem.

It is also uniformly continuous on R\mathbb R. These properties depend on continuity; they do not follow from periodicity alone.

Differentiation

If ff is differentiable and has period TT, differentiating its defining identity gives

f′(x+T)=f′(x).f'(x+T)=f'(x).

Thus its derivative is periodic with period TT, although its least period need not be TT, and it may be constant.

The converse is not automatic. If f′f' has period TT, then

f(x+T)−f(x)f(x+T)-f(x)

is constant in xx, but that constant need not be zero. For example, the derivative of f(x)=xf(x)=x is periodic, while ff itself is not.

Integration and mean value

For a locally integrable TT-periodic function, the integral over a complete period does not depend on the starting point:

∫aa+Tf(x) dx=∫0Tf(x) dx.\int_a^{a+T}f(x)\,dx=\int_0^T f(x)\,dx.

Its mean over a period is

f‾=1T∫0Tf(x) dx.\overline f=\frac1T\int_0^T f(x)\,dx.

If

F(x)=∫0xf(t) dt,F(x)=\int_0^x f(t)\,dt,

then

F(x+T)−F(x)=Tf‾.F(x+T)-F(x)=T\overline f.

Therefore, FF is TT-periodic exactly when ff has zero mean. In general,

F(x)=f‾ x+p(x),F(x)=\overline f\,x+p(x),

where pp is TT-periodic. Integration thus separates a periodic input into a linear drift and a periodic remainder.

Fourier representation

Fourier series express periodic functions in terms of harmonically related sinusoids or complex exponentials. For a sufficiently regular function with period TT,

f(x)∼∑n∈Zcne2πinx/T,f(x)\sim\sum_{n\in\mathbb Z}c_n e^{2\pi i n x/T},

where

cn=1T∫0Tf(x)e−2πinx/T dx.c_n=\frac1T\int_0^T f(x)e^{-2\pi i n x/T}\,dx.

The coefficient c0c_0 is the mean value. The other coefficients describe components at integer multiples of the base frequency 1/T1/T. The specified period TT need not be the fundamental period.

The meaning of the expansion depends on regularity and on the type of convergence:

  • For a square-integrable function, the partial sums converge in the L2L^2 sense.
  • For a piecewise continuously differentiable periodic function, the series converges at each point to the average of the left and right limits.
  • At a continuity point under those conditions, it converges to the function’s value.
  • Near jump discontinuities, partial sums exhibit persistent overshoot known as the Gibbs phenomenon.

Periodicity alone does not guarantee pointwise convergence of a Fourier series.

Fourier methods developed centrally through Joseph Fourier’s work on heat conduction in the early nineteenth century, building on earlier investigations of vibrating strings and trigonometric expansions. They became foundational tools in harmonic analysis and in solving partial differential equations.

Discrete periodicity and sampling

For a sequence indexed by integers, periodicity means that there is a positive integer NN such that

xn+N=xn(n∈Z).x_{n+N}=x_n\qquad(n\in\mathbb Z).

Unlike a real-variable function, every periodic sequence has a least positive integer period, because every nonempty set of positive integers has a least element.

The sequence

xn=eiωnx_n=e^{i\omega n}

is periodic exactly when

ω2π∈Q.\frac{\omega}{2\pi}\in\mathbb Q.

If ω/(2π)=p/q\omega/(2\pi)=p/q in lowest terms with q>0q>0, its fundamental period is qq.

Sampling a continuous periodic function does not always produce a periodic sequence. If ff has period TT and is sampled at intervals Δ\Delta, then Δ/T=p/q\Delta/T=p/q guarantees that the sampled sequence repeats after qq steps, though its fundamental period may be smaller. An irrational sampling ratio does not provide such a guarantee.

The discrete Fourier transform represents a finite data block using discrete harmonics and corresponds to a periodic extension of that block. This mathematical extension need not reflect actual repetition in the original measured signal.

Related forms of recurrence

Exact periodicity should be distinguished from several broader notions:

  • Eventual periodicity: exact repetition begins only after an initial interval or finite prefix.
  • Quasiperiodicity: motion or variation is built from finitely many independent basic frequencies, typically with irrational frequency ratios, and need not repeat exactly.
  • Almost periodicity: in a standard uniform formulation, arbitrarily accurate approximate repetitions occur with relatively dense translation shifts.
  • Antiperiodicity: a relation f(x+T)=−f(x)f(x+T)=-f(x). It implies f(x+2T)=f(x)f(x+2T)=f(x), so 2T2T is a period.

In several variables, a function may be invariant under several independent translation vectors. In complex analysis, nonconstant meromorphic functions with two periods independent over the real numbers are called elliptic functions. A nonconstant function holomorphic on the entire complex plane cannot have two such independent periods: double periodicity would make it bounded, forcing it to be constant.

Applications and modeling limits

Periodic functions describe idealized alternating electrical signals, acoustic tones, rotational motion, and traveling or standing waves. Periodic solutions of a differential equation represent trajectories that return exactly to the same state after a fixed interval. Periodic boundary conditions identify opposite boundaries of a computational or physical domain, allowing a repeating cell to represent a larger structure.

In time-series analysis, seasonal effects are often modeled with periodic functions. Observed cycles, however, may change in amplitude, phase, or duration. Repeated peaks or a strong spectral component therefore do not by themselves establish exact periodicity. A finite set of observations also cannot uniquely determine whether the underlying function repeats for all arguments.