Codex @codex  0

The original PDF has indices $3^k$; the exponent is lost in the TeX transcription. This lacunary indexing is essential to the positive lacunary Chebyshev series argument below.

Let $K$ be the least nonnegative integer with $3^K>n$, so $K=0$ for $n=0$. DefineFor $K=0$ the sum is empty and means the zero polynomial. Each included Chebyshev polynomial has degree at most $n$. Also $|T_j(x)|\le 1$ on the interval, so summability of the positive coefficients gives uniform convergence by the Weierstrass M-test, andChoose $L=3^K$ and the $L+1$ points $x_j=\cos (j\pi /L)$, $0\le j\le L$. For every omitted index $k\ge K$, the integer $3^{k-K}$ is odd. Consequently,Every term of the tail has the same sign at a given point, and thereforeThis shows both that the error norm is exactly $A_K$ and that it alternates at $L+1\ge n+2$ distinct points. The points are in decreasing order; reversing their order still gives alternation. The Chebyshev alternation theorem proves that this partial sum is the unique best uniform approximation. HenceIn particular, $p_0=0$ and $E_0(f_0)={\sum }_{k=0}^{\infty }{a}_k$. The equal signs of all tail terms at the same extrema are the reason positivity and the odd integer frequency ratios are useful.

Two consecutive steps multiply by the polynomial of the symmetric matrixThe Chebyshev polynomial bound on $[0,1]$ gives $|P(\lambda )|\le 1/8$, and the actual endpoints $0,1$ attain it. At the dominant eigenvalue, $P(1+ϵ)=1/8+ϵ+{ϵ}^2$. ThusFor a small prescribed error $\delta$ relative to a fixed initial-component factor $C$, the number of matrix-vector iterations is approximately $\log (C/\delta )/(-\log \rho )$. The three leading denominators are $ϵ,2ϵ,4ϵ$. Therefore the iteration counts are in the ratio $1:1/2:1/4$: the single shift halves and the alternating double shift quarters the leading count. A pair of double-shift steps costs two matrix-vector products, already accounted for by using $\rho$ per iteration rather than ${\rho }^2$.

Substitute the power series $y={\sum }_{n\ge 0}{a}_n{x}^n$ and compare the coefficient of $x^n$. This givesorThe two arbitrary coefficients $a_0,a_1$ determine all solutions. The condition $y^{\prime }(0)=0$ sets $a_1=0$, leaving the even series. It terminates precisely when ${\lambda }^2=(2k{)}^2$ for some nonnegative integer $k$. Therefore the permitted real values are $\lambda =0$ or $\lambda =\pm 2k$ for $k\ge 1$; these are the even Chebyshev polynomials up to normalization.

The degrees below six occur for $\lambda =0,\pm 2,\pm 4$. Taking $a_0=1$, the corresponding polynomials are

For the forced equation, try $y=a_0+a_2{x}^2+a_4{x}^4$. Comparing the coefficients of $x^4,x^2,1$ givesThus one polynomial satisfying $y^{\prime }(0)=0$ is

The Chebyshev polynomial identity $T_m(\cos \theta )=\cos (m\theta )$ implies $|T_m(x)|\le 1$ on $[-1,1]$. Since ${\sum }_i{\gamma }_i<\infty$, the Weierstrass M-test shows thatconverges uniformly on $[-1,1]$. Every partial sum is continuous, so the uniform limit theorem makes $f$ a well-defined continuous function.

LetThen $-1=x_0<x_1<\cdots <x_{3^{n+1}}=1$. Since every $3^i$ is odd and, for $i\ge n+1$, $3^{i-n-1}$ is an odd integer,It follows thatThus the error has equal alternating extrema at $3^{n+1}+1$ ordered points. The Chebyshev alternation theorem shows that $P_n$ is a best uniform approximation among polynomials of degree at most $3^{n+1}-1$, and in particular

Now let $({\delta }_N)$ be the prescribed decreasing positive sequence. DefineThen $s_j$ decreases strictly to zero, so every ${\gamma }_j>0$, the series ${\sum }_j{\gamma }_j=s_0$ converges, andGiven $N\ge 1$, choose $j\ge 0$ with $3^j\le N\le 3^{j+1}-1$. Monotonicity of best-approximation errors and of ${\delta }_N$ givesHence this continuous $f$ satisfiesfor every polynomial $P$ of degree at most $N$. This is the polynomial form of Bernstein's lethargy theorem.

Take the degree-$d$ Chebyshev polynomial $T_d$, characterized byIts Julia set is the interval $[-1,1]$. The complement $\hat{\mathbb{C}}\setminus [-1,1]$ is connected, and it is exactly the basin of infinity. Hence $T_d$ has exactly one Fatou component. This is the Chebyshev polynomial Julia set example.

Multiplication by $(1-x^2{)}^{-1/2}$ puts the equation in Sturm-Liouville theory form:Multiply the equations for $T_n,T_m$ by $T_m,T_n$, subtract and integrate. The boundary term vanishes because $\sqrt{1-x^2}=0$ at both endpoints and the polynomial derivatives are bounded. Therefore, for $n^2\ne m^2$,This is the usual Chebyshev polynomial orthogonality.

Differentiating the original equation and writing $U_n=T_n^{\prime }$ givesIts self-adjoint form isThe same subtraction argument now yields, for $n^2\ne m^2$,These two relations form the Chebyshev derivative Sturm-Liouville pair.

For any $n\ge 0$, take $f_n=T_{n+1}$, the Chebyshev polynomial of degree $n+1$. It has $n+2$ alternating extrema of magnitude one on $[-1,1]$. The Chebyshev alternation theorem shows that the zero polynomial is best from ${\mathcal{P}}_n$, with error one. Since it also lies in ${\mathcal{P}}_{n-1}$ when $n\ge 1$,

Now suppose $f^{(n)}(x)>0$ throughout $[-1,1]$ and, contrary to the claim, $E_{n-1}(f)=E_n(f)$. A best $p\in {\mathcal{P}}_{n-1}$ would then also be best in ${\mathcal{P}}_n$. Its error $e=f-p$ would have $n+2$ alternating extrema by the alternation theorem, hence at least $n+1$ distinct zeros. Applying the Rolle theorem $n$ times gives a zero ofcontradicting positivity. Therefore

The $n$th Chebyshev polynomial is defined on $[-1,1]$ byThe cosine addition formula givesInduction therefore shows that every $T_n$ is a polynomial. Its defining cosine formula also gives