Increments and finite differences of functions

[11/91%]
Show
LaTeX

For ff a real valued function on the real line, define the function Δf\Delta f by Δf(x)=f(x+1)−f(x)\Delta f(x) = f(x+1)-f(x). For n≥2n \geq 2, define Δnf\Delta^{n} f recursively by Δnf=Δ(Δn−1f)\Delta^{n} f = \Delta (\Delta^{n-1} f). Prove that Δnf=0\Delta^{n} f = 0 if and only if ff has the form f(x)=a0(x)+a1(x)x+⋯+an−1(x)xn−1f(x) = a_{0}(x)+a_{1}(x)x+\cdots +a_{n-1}(x)x^{n-1} where a0,a1,…,an−1a_{0}, a_{1}, \ldots , a_{n-1} are periodic functions of period 1.

?

Let

△f(x)=f(x+△x)−f(x) \triangle f(x)=f(x+\triangle x)-f(x)

and

△2f(x)=△{△f(x)} \triangle ^{2}f(x)=\triangle \left\{ \triangle f(x)\right\}

be, respectively, the first- and second-order finite differences of the function f(x)f(x).

Prove that if the function f(x)f(x) (−∞<x<+∞-\infty <x<+\infty) is continuous and

△2f(x)≡0, \triangle ^{2}f(x)\equiv 0 ,

then this function is linear, i.e.,

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

where aa and bb are constants.

?

If xx changes from 11 to 10001000, find the increment Δx\Delta x of the independent variable xx and the corresponding increment Δy\Delta y of the function y=lg⁡xy=\lg x.

?

If xx changes from 0.010.01 to 0.0010.001, find the increment Δx\Delta x of the independent variable xx and the corresponding increment Δy\Delta y of the function y=1x2y=\frac{1}{x^{2}}.

?

Let:

?
(a)

y=ax+by=ax+b;

(b)

y=ax2+bx+cy=ax^{2}+bx+c;

(c)

y=axy=a^{x}.

If the variable xx receives an increment Δx\Delta x, find the increment Δy\Delta y.

Prove that:

?
(a)

Δ[f(x)+g(x)]=Δf(x)+Δg(x)\Delta \left[f(x)+g(x)\right]=\Delta f(x)+\Delta g(x);

(b)

Δ[f(x)g(x)]=g(x+Δx)Δf(x)+f(x)Δg(x)\Delta \left[f(x)g(x)\right]=g(x+\Delta x)\Delta f(x)+f(x)\Delta g(x).

For the function y=x3−3x2+5x−7y = x^{3} - 3x^{2} + 5x - 7 at x=0x=0, the values of the differences Δy0=+3\Delta y_{0} = +3, Δ2y0=0\Delta^{2} y_{0} = 0, Δ3y0=5\Delta^{3} y_{0} = 5 have been computed, corresponding to an increment h0=1h_{0}=1 of the argument. It is required to compute the differences Δy0‾\overline{\Delta y_{0}}, Δ2y0‾\overline{\Delta^{2} y_{0}}, Δ3y0‾\overline{\Delta^{3} y_{0}}, corresponding to an increment of the argument h1=110h_{1} = \dfrac {1}{10}.

?

For the function y=2x4−x3+3x2−1y = 2x^{4} - x^{3} + 3x^{2} - 1 at x=0x=0, the differences Δy0=4\Delta y_{0} = 4, Δ2y0=28\Delta^{2} y_{0} = 28, Δ3y0=66\Delta^{3} y_{0} = 66, Δ4y0=48\Delta^{4} y_{0} = 48 have been computed, corresponding to an increment h0=1h_{0}=1 of the argument. It is required to compute the differences Δy0‾\overline{\Delta y_{0}}, Δ2y0‾\overline{\Delta^{2} y_{0}}, Δ3y0‾\overline{\Delta^{3} y_{0}}, Δ4y0‾\overline{\Delta^{4} y_{0}}, corresponding to an increment of the argument h1=110h_{1} = \dfrac {1}{10}.

?

Prove that if the function ff is mm times differentiable on the interval (a,b)(a, b) and x,x+mh∈(a,b)x, x+m h \in (a, b), then

Δhmf(x)=hmf(m)(x+θh), \Delta _{h}^{m} f(x)=h^{m} f^{(m)}(x+\theta h),

where 0<θ<m0<\theta <m (for the definition of Δhmf(x)\Delta_{h}^{m} f(x) see problem III.3.1).

?

Let the function ff be continuous on R\mathbb {R}, and suppose there exist real numbers ak,bk,ck(1⩽k⩽n)a_{k}, b_{k}, c_{k}(1 \leqslant k \leqslant n) such that

f(x)=∑1⩽k⩽nakf(bkx+cky) for all x,y∈R. f(x)=\sum _{1 \leqslant k \leqslant n} a_{k} f\left(b_{k} x+c_{k} y\right) \quad \text{ for all } \quad x, y \in \mathbb {R} .

Prove that if ck≠0c_{k} \neq 0 for all kk, then ff is a polynomial.

?

Let n∈N,f:[0,1]→Rn \in \mathbb {N}, f:\left[0,1\right] \rightarrow \mathbb {R}. The polynomial

Bn(f,x)=∑0⩽k⩽nCnkf(kn)xk(1−x)n−k B_{n}(f, x)=\sum _{0 \leqslant k \leqslant n} C_{n}^{k} f\left(\frac{k}{n}\right) x^{k}(1-x)^{n-k}

is called the Bernstein polynomial of the function ff.

Let the function ff be defined on the interval [0,1]\left[0,1\right]. Prove that

Bn(f,x)=f(0)+∑1⩽k⩽nxkCnkΔknf(0), B_{n}(f, x)=f(0)+\sum _{1 \leqslant k \leqslant n} x^{k} C_{n}^{k} \Delta _{k}^{n} f(0),

where Δ1nf(y)=f(y+1n)−f(y),Δk+1nf(y)=Δ1n(Δknf(y)),k∈N\quad \Delta_{1}^{n} f(y)=f\left(y+\frac{1}{n}\right)-f(y), \quad \Delta_{k+1}^{n} f(y)=\Delta_{1}^{n}\left(\Delta_{k}^{n} f(y)\right), \quad k \in \mathbb {N}. In particular, if ff is a polynomial, then the degree of Bn(f)B_{n}(f) does not exceed the degree of ff.

?

References

[deSi04] de Souza P.N., Silva J. Berkeley Problems in Mathematics. 3rd ed., Springer, 2004.
[Dem97] Demidovich B.P. Сборник задач и упражнений по математическому анализу. 13-е изд., Издательство Московского университета / Издательство "ЧеРо", 1997.
[ГюКу03] Гюнтер Н.М., Кузьмин Р.О. Сборник задач по высшей математике. 13-е изд., Издательство «Лань», 2003.
[МГЛ+04] Макаров Б.М., Голузина М.Г., Лодкин А.А., Подкорытов А.Н. Избранные задачи по вещественному анализу. 2-е изд., Невский Диалект; БХВ-Петербург, 2004.