ApproxFun.jl Documentation

ApproxFun is a package for approximating and manipulating functions, and for solving differential and integral equations.


A basic approach of computational mathematics that ApproxFun exploits is expansion in a basis

\[\mathop{f}(x) \approx \sum_{k=1}^n c_k \mathop{ψ}_k(x).\]

Some traditional examples of bases $\mathop{ψ}_1(x), \mathop{ψ}_2(x), …$ are

  1. Taylor series: $1, x, x^2, …$
  2. Fourier series (for periodic functions on 0..2π): $1, \sin{x}, \cos{x}, \sin{2x}, …$
  3. Chebyshev series (for non-periodic functions on -1..1): $1, x, \cos(2\arccos{x}), \cos(3\arccos{x}), …$

In ApproxFun, functions are represented by a Fun with two components: space, which dictates the basis and coefficients which is a finite vector of coefficients. Note that each Fun can have a different length vector of coefficients, allowing for approximation of many different functions to high accuracy.

The approximation by a Fun can be determined by a variety of methods:

(1) Explicitly specifying the coefficients:

julia> f = Fun(Taylor(),[1,2,3])  # Represents 1 + 2x + 3x^2Fun(Taylor(🕒), [1, 2, 3])
julia> f(1.0)6.0

(2) Constructors take in a Function and adaptively determine the number of coefficients. For example,

julia> Fun(exp)Fun(Chebyshev(), [1.26607, 1.13032, 0.271495, 0.0443368, 0.00547424, 0.000542926, 4.49773e-5, 3.19844e-6, 1.99212e-7, 1.10368e-8, 5.5059e-10, 2.49796e-11, 1.03911e-12, 3.99368e-14])

determines that f can be approximated to roughly machine precision using 14 coefficients. See Constructors for more information.

(3) Manipulation of Funs give new Funs, where the number of coefficients is determined from the input. The simplest example is addition, which for compatible bases is just padding the vectors to the same length and adding.

julia> a = Fun(cos,Chebyshev()); ncoefficients(a)13
julia> b = Fun(x->cos(10cos(x^2)),Chebyshev()); ncoefficients(b)64
julia> ncoefficients(a+b)64

On the other hand, multiplication results in an approximation with more coefficients than either a or b, so that the result approximates the true a*b to roughly machine accuracy:

julia> ncoefficients(a*b)76
julia> a(0.1)*b(0.1) - (a*b)(0.1)1.1102230246251565e-16

The example of multiplication highlights the importance of adaptivity: if with a fixed discretization size, operations like multiplication would lose accuracy when the true function is no longer resolved by the discretization. More complicated examples are solving differential equations, where the coefficients of the solution can be determined adaptively, see Equations.

ApproxFun supports a number of different spaces, as described in Spaces. A key component of ApproxFun is support for interaction between different spaces. This is crucial for efficient solution of differential equations, where linear operators are described as acting between different spaces, see Operators.