An introduction is available in this video from JuliaCon 2018.
Taylor models provide a way to rigorously manipulate and evaluate functions using floating-point arithmetic. They have been widely used for validated computing: in global optimization and range bounding, for validated solutions of ODEs, rigorous quadrature, etc.
A Taylor model (TM) of order $n$ for a function $f$ which is supposed to be $n + 1$ times continuously differentiable over an interval $[a,b]$, is a rigorous polynomial approximation of $f$. Specifically, it is a couple $(P, \Delta)$ formed by a polynomial $P$ of degree $n$, and an interval part $\Delta$, such that $f(x) − P(x) \in \Delta$, $\forall x ∈ [a,b]$. Roughly speaking, as their name suggests, the polynomial can be seen as a Taylor expansion of the function at a given point. The interval $\Delta$ (also called interval remainder) provides the validation of the approximation, meaning that it provides an enclosure of all the approximation errors encountered (truncation, roundings).
Here we generate TMs of order 6 and 7 over $I = [-0.5,1.0]$. We can view a TM as a a tube around the actual function.
using TaylorModels f(x) = x*(x-1.1)*(x+2)*(x+2.2)*(x+2.5)*(x+3)*sin(1.7*x+0.5) a = -0.5 .. 1.0 # Domain x0 = mid(a) # Expansion point tm6 = TaylorModel1(6, interval(x0), a) # Independent variable for Taylor models, order 6 tm7 = TaylorModel1(7, interval(x0), a) # Independent variable for Taylor models, order 7 # Taylor models corresponding to f(x) of order 6 and 7 ftm6 = f(tm6) ftm7 = f(tm7) # Now the plot using Plots; gr() plot(range(inf(a), stop=sup(a), length=1000), x->f(x), label="f(x)", lw=2, xaxis="x", yaxis="f(x)") plot!(ftm6, label="6th order") plot!(ftm7, label="7th order")
- Luis Benet, Instituto de Ciencias Físicas, Universidad Nacional Autónoma de México (UNAM)
- David P. Sanders, Departamento de Física, Facultad de Ciencias, Universidad Nacional Autónoma de México (UNAM)
- Rigorous Polynomial Approximations and Applications, Mioara Maria Joldes, Ecole normale supérieure de lyon - ENS LYON (2011)
Financial support is acknowledged from DGAPA-UNAM PAPIIT grants IN-117117, IG-100616 and IG-100819. DPS acknowledges support through a Cátedra Marcos Moshinsky (2018).