阶的估计I 无穷小量与强函数2 Taylor公式 基本初等函数与三角函数的阶

阶的估计I 无穷小量与强函数2 Taylor公式 基本初等函数与三角函数的阶

这一讲介绍Taylor公式在阶的估计中的应用,并基于Taylor公式给出一些常用函数的阶的估计。


定理1.2 Taylor公式
x 0 x_0 x0的某个领域内,如果 f ( x ) f(x) f(x)的前 n n n阶导存在,且 f ( n ) ( x ) = O ( 1 ) f^{(n)}(x)=O(1) f(n)(x)=O(1),则
f ( x ) = ∑ k = 0 n − 1 f ( k ) ( x 0 ) k ! ( x − x 0 ) k + O ( ∣ x − x 0 ∣ n ) = ∑ k = 0 n f ( k ) ( x 0 ) k ! ( x − x 0 ) k + o ( ∣ x − x 0 ∣ n + 1 ) f(x) = \sum_{k=0}^{n-1} \frac{f^{(k)}(x_0)}{k!}(x-x_0)^k+O(|x-x_0|^n) \\ = \sum_{k=0}^{n} \frac{f^{(k)}(x_0)}{k!}(x-x_0)^k+o(|x-x_0|^{n+1}) f(x)=k=0n1k!f(k)(x0)(xx0)k+O(xx0n)=k=0nk!f(k)(x0)(xx0)k+o(xx0n+1)

评注 下面是一些常用的估计式,当 x → 0 x \to 0 x0时,
sin ⁡ x = x + o ( x ) = x − x 3 3 ! + o ( x 3 ) = x − x 3 3 ! + O ( x 5 ) cos ⁡ x = 1 − x 2 2 + o ( x 2 ) = 1 − x 2 2 + O ( x 4 ) log ⁡ ( 1 + x ) = x + o ( x ) = x − x 2 2 + O ( x 3 ) ( 1 + x ) c = 1 + c x + o ( x ) = 1 + c x + O ( x 2 ) e x = 1 + x + o ( x ) = 1 + x + x 2 2 ! + O ( x 3 ) \sin x = x+o(x)=x-\frac{x^3}{3!}+o(x^3)=x-\frac{x^3}{3!}+O(x^5) \\ \cos x = 1-\frac{x^2}{2}+o(x^2) = 1-\frac{x^2}{2}+O(x^4) \\ \log(1+x) = x+o(x) = x-\frac{x^2}{2}+O(x^3) \\ (1+x)^c = 1+cx+o(x) = 1+cx + O(x^2) \\ e^x=1+x +o(x)=1+x+\frac{x^2}{2!}+O(x^3) sinx=x+o(x)=x3!x3+o(x3)=x3!x3+O(x5)cosx=12x2+o(x2)=12x2+O(x4)log(1+x)=x+o(x)=x2x2+O(x3)(1+x)c=1+cx+o(x)=1+cx+O(x2)ex=1+x+o(x)=1+x+2!x2+O(x3)

更一般地,如果 f ( x ) = o ( 1 ) , x → 0 f(x)=o(1),x \to 0 f(x)=o(1),x0,则上面的所有 x x x都可以换成 f ( x ) f(x) f(x);另外, e x e^x ex的估计在 ∣ x ∣ < 1 |x|<1 x<1时成立, cos ⁡ ( x ) \cos(x) cos(x)的估计在 ∣ x ∣ < ∞ |x|<\infty x<时成立。


例1.2 用Taylor公式计算数列极限
lim ⁡ n → ∞ n ( n 1 / n − 1 ) log ⁡ n \lim_{n \to \infty} \frac{n(n^{1/n}-1)}{\log n} nlimlognn(n1/n1)

我们先分析 n 1 / n n^{1/n} n1/n,当 n → ∞ n \to \infty n时, 1 n log ⁡ n = o ( 1 ) \frac{1}{n}\log n = o(1) n1logn=o(1),用Taylor公式,
n 1 / n = e 1 n log ⁡ n = 1 + 1 n log ⁡ n + O ( log ⁡ 2 n n 2 ) n^{1/n}=e^{\frac{1}{n}\log n} = 1+\frac{1}{n}\log n +O\left( \frac{\log^2 n}{n^2} \right) n1/n=en1logn=1+n1logn+O(n2log2n)

所以
n ( n 1 / n − 1 ) log ⁡ n = log ⁡ n + O ( log ⁡ 2 n n ) log ⁡ n = 1 + O ( log ⁡ n n ) = 1 + o ( 1 ) \frac{n(n^{1/n}-1)}{\log n} = \frac{\log n+O\left( \frac{\log^2 n}{n} \right)}{\log n} = 1+O\left( \frac{\log n}{n}\right) = 1+o(1) lognn(n1/n1)=lognlogn+O(nlog2n)=1+O(nlogn)=1+o(1)

对无穷小量与强函数的运算法则不熟悉的可以参考上一讲;第一个等号中我们使用了强函数可积的性质(注意函数一定是自身的强函数)
n O ( log ⁡ 2 n n 2 ) = O ( n ) O ( log ⁡ 2 n n 2 ) = O ( log ⁡ 2 n n ) nO\left( \frac{\log^2 n}{n^2} \right)=O(n)O\left( \frac{\log^2 n}{n^2} \right) = O\left( \frac{\log^2 n}{n} \right) nO(n2log2n)=O(n)O(n2log2n)=O(nlog2n)

第二个等号的计算与第一个等号类似;第三个等号用到的是第三个法则,某函数的强函数是无穷小量,则它本身也是无穷小量,因为 1 n log ⁡ n = o ( 1 ) \frac{1}{n}\log n = o(1) n1logn=o(1),所以 O ( log ⁡ n n ) = o ( 1 ) O\left( \frac{\log n}{n}\right)=o(1) O(nlogn)=o(1);因此
lim ⁡ n → ∞ n ( n 1 / n − 1 ) log ⁡ n = 1 \lim_{n \to \infty} \frac{n(n^{1/n}-1)}{\log n}=1 nlimlognn(n1/n1)=1

例1.3 用Taylor公式得到数列的近似
假设 α = O ( 1 ) , β = O ( 1 ) \alpha=O(1),\beta=O(1) α=O(1),β=O(1),则
( n + α ) n + β = n n + β e α [ 1 + α ( β − α 2 ) n + O ( n − 2 ) ] (n+\alpha)^{n+\beta}=n^{n+\beta}e^{\alpha}\left[ 1+\frac{\alpha(\beta-\frac{\alpha}{2})}{n} +O(n^{-2})\right] (n+α)n+β=nn+βeα[1+nα(β2α)+O(n2)]

计算 ( n + α ) n + β (n+\alpha)^{n+\beta} (n+α)n+β
log ⁡ ( n + α ) n + β = ( n + β ) log ⁡ ( n + α ) = ( n + β ) [ log ⁡ n + log ⁡ ( 1 + α n ) ] = ( n + β ) [ log ⁡ n + α n − α 2 2 n 2 + O ( α 3 n 3 ) ] = ( n + β ) log ⁡ n + α + α ( β − α 2 ) 1 n + O ( 1 n 2 ) \log(n+\alpha)^{n+\beta} = (n+\beta)\log(n+\alpha) \\ = (n+\beta)\left[ \log n+\log(1+\frac{\alpha}{n}) \right] \\ =(n+\beta)\left[ \log n+\frac{\alpha}{n}-\frac{\alpha^2}{2n^2}+O(\frac{\alpha^3}{n^3})\right] \\ = (n+\beta)\log n+\alpha+\alpha(\beta-\frac{\alpha}{2})\frac{1}{n}+O(\frac{1}{n^2}) log(n+α)n+β=(n+β)log(n+α)=(n+β)[logn+log(1+nα)]=(n+β)[logn+nα2n2α2+O(n3α3)]=(n+β)logn+α+α(β2α)n1+O(n21)

所以
( n + α ) n + β = n n + β e α [ 1 + α ( β − α 2 ) n + O ( n − 2 ) ] (n+\alpha)^{n+\beta} = n^{n+\beta}e^{\alpha}\left[ 1+\frac{\alpha(\beta-\frac{\alpha}{2})}{n} +O(n^{-2})\right] (n+α)n+β=nn+βeα[1+nα(β2α)+O(n2)]

基于这个结果我们可以得到一个常用极限:
( 1 + α n ) n → e α , n → ∞ (1+\frac{\alpha}{n})^n \to e^{\alpha},n \to \infty (1+nα)neα,n

上式左右除以 n n + β n^{n+\beta} nn+β
( 1 + α n ) n + β = e α [ 1 + α ( β − α 2 ) n + O ( n − 2 ) ] (1+\frac{\alpha}{n})^{n+\beta}=e^{\alpha}\left[ 1+\frac{\alpha(\beta-\frac{\alpha}{2})}{n} +O(n^{-2})\right] (1+nα)n+β=eα[1+nα(β2α)+O(n2)]

β = 0 \beta=0 β=0,则
( 1 + α / n ) n = e α [ 1 − α 2 2 n + O ( n − 2 ) ] (1+\alpha/n)^n=e^{\alpha}\left[ 1-\frac{\alpha^2}{2n} +O(n^{-2})\right] (1+α/n)n=eα[12nα2+O(n2)]

n → ∞ n \to \infty n即可。

猜你喜欢

转载自blog.csdn.net/weixin_44207974/article/details/112424081