Threads
DFALG 2Vector spaces
Given n distinct inputs t1,…,tn and prescribed outputs y1,…,yn, polynomial interpolation finds the polynomial f∈Pn−1 satisfying
f(ti)=yi,i=1,…,n.Using the standard basis
B=(1,t,t2,…,tn−1),write
f(t)=a0+a1t+⋯+an−1tn−1.The coefficients form the coordinate vector
[f]B=a0a1⋮an−1.- 1Coordinates and linear structure / coordinates of a polynomial
- 4Local approximation / match values instead of derivatives
DFCAL 3Taylor series, sequences and infinite series
If f has derivatives through order n at a, its Taylor polynomial of degree at most n about a is
Pn(x)=k=0∑nk!f(k)(a)(x−a)k.- 4Local approximation / one variable: match derivatives at a centre
THCAL 3Taylor series, sequences and infinite series
Suppose f has n+1 continuous derivatives on an interval containing a and x. If
Pn(x)=k=0∑nk!f(k)(a)(x−a)k,then
f(x)=Pn(x)+Rn+1(x),where the integral remainder is
Rn+1(x)=n!1∫axf(n+1)(t)(x−t)ndt.For x=a, there is also a point c between a and x for which the Lagrange remainder is
Rn+1(x)=(n+1)!f(n+1)(c)(x−a)n+1.- 4Local approximation / the error term
THCAL 3Taylor series, sequences and infinite series
Suppose k is an integer with k≥2 and f has k+1 continuous derivatives on a neighbourhood of a. If
f(j)(a)=0(1≤j<k),f(k)(a)=0.then Taylor's theorem gives
f(x)−f(a)=k!f(k)(a)(x−a)k+Rk+1(x),Rk+1(x)=O(∣x−a∣k+1),Rk+1(x)=o(∣x−a∣k).Thus, locally:
- if k is even and f(k)(a)>0, a is a local minimum;
- if k is even and f(k)(a)<0, a is a local maximum;
- if k is odd, the leading term changes sign across a, giving a horizontal point of inflection.
- 4Local approximation / classifying stationary points in one variable
THCAL 4Functions of several variables
If f is differentiable at a=(x0,y0) and z0=f(x0,y0), then the tangent plane to z=f(x,y) at (x0,y0,z0) is
z=f(x0,y0)+fx(x0,y0)(x−x0)+fy(x0,y0)(y−y0).- 4Local approximation / first-order model in several variables
DFCAL 4Functions of several variables
Let a=(x0,y0), hx=x−x0 and hy=y−y0. If f has continuous partial derivatives through order three near a, its Taylor polynomial through total degree three is
T3(x,y)=f(a)+fx(a)hx+fy(a)hy+2!1(fxx(a)hx2+2fxy(a)hxhy+fyy(a)hy2)+3!1(fxxx(a)hx3+3fxxy(a)hx2hy+3fxyy(a)hxhy2+fyyy(a)hy3).The degree-two part can be written with the Hessian as
T2(x)=f(a)+∇f(a)⋅(x−a)+21(x−a)THf(a)(x−a).- 4Local approximation / second-order model: the Hessian
THCAL 4Functions of several variables
Let x0=(x0,y0) be a stationary point. Assume that f has continuous second-order partial derivatives on a neighbourhood of x0 (so f is C2 there). Define
D=detHf(x0)=fxx(x0)fyy(x0)−fxy(x0)2.Then:
- if D>0 and fxx(x0)>0, x0 is a local minimum;
- if D>0 and fxx(x0)<0, x0 is a local maximum;
- if D<0, x0 is a saddle point;
- if D=0, the test is inconclusive.
- 4Local approximation / classifying stationary points in several variables
4 / Algebra / Calculus / 7 stops
Replace a hard object by a simple model that matches it near a point, and bound the error.
- Within a chapter
- Waiting between chapters
- A stop
- One stop on two threads
- Pinned
- The thread in question