当前位置:首页|资讯

泰勒公式的应用(习题篇)

作者:现代微积分发布时间:2024-09-12

前言

最近任务又开始多起来了,趁现在学业还不算太繁忙,加急将这篇文章写好,然后极限板块的内容就先告一段落了。其实再往后还可以继续串连知识的,比如泰勒中值定理(带拉格朗日余项的泰勒公式,可以定量地估计误差),泰勒级数等,再往后还可跟函数项级数串一起讨论。而这又得花不少的时间去深化了解既有的知识了,总之秉持"进一寸有一寸的欢喜"的学习状态就好了~

另外,由于小破站有公式条数限制,因此有些公式就用了例如√3,2/5,x²这种可输入的写法了,也请将就一下~

补充知识

无穷远处的泰勒展开

前两篇文章

泰勒公式的证明和应用(基础篇)

泰勒公式间接展开法(拔高篇)

分别介绍了泰勒展开的直接法和间接法,第二篇文章文末有总结4个选阶准则,分别用于处理函数加减和数乘(准则1)函数乘法(准则2)函数复合(准则3)函数的导函数/原函数(准则4)下文提到准则(1)(2)(3)(4)时就默认指定这几个准则了,新来的观众可以先看完这两篇文章哦

其中复合函数的泰勒展开定理简述如下:

已知函数f(x)在x=x₀处n阶可导,且当t→t₀时函数g(t)→x₀。

若把内层函数g(t)简记为方框□,则复合函数f(□)可作如下的展开:

%5CBox%20%5Cto%20x_0%2Cf(%5CBox%20)%3D%5Csum_%7Bk%3D0%7D%5E%7Bn%7D%5Cfrac%7Bg%5E%7B(k)%7D(x_0)%7D%7Bk!%7D(%5CBox-x_0)%5Ek%2B%5Comicron%20((%5CBox-x_0)%5En)

例如:

熟知广义二项式定理:

%5Cbbox%5B%238fffa8%5D%7B(1%2Bx)%5E%5Calpha%20%3D1%2B%5Calpha%20x%2B%5Cfrac%7B%5Calpha%20(%5Calpha%20-1)%7D%7B2!%7Dx%5E2%2B%5Ccdots%20%2BC_%5Calpha%20%5Enx%5En%2B%5Comicron%20(x%5En)%20%7D

这个在第一篇文章中用直接法推导过哦

以α取1/2时的前三项为例:

√(1+x)=1+x/2-x²/8+o(x²)

由复合函数的泰勒展开定理,这里的x可替换为一个函数值趋于0的函数,例如:

x→0时x²→0,于是就有:

当x→0时,

√(1+)=1+()/2-()²/8+o(()²)

=1+x²/2-x⁴/8+o(x⁴)


这里还是得再提一嘴细节,这里的o((x²)²)是用第一复合法则证明的,并不是直接把o(x²)里面的·x换为x²哦,实际过程如下:

已知x→0,√(1+x)=1+x/2-x²/8+o(x²),

那么[√(1+x)-(1+x/2-x²/8)]/x²极限为0

进而由第一复合法则得:

[√(1+)-(1+()/2-()²/8)]/(极限为0

⇒√(1+)-(1+()/2-()²/8)=o(()²)

移项即得√(1+)=(1+()/2-()²/8)+o(()²)

一般的复合函数泰勒展开原理也是这样,在前一篇文章(间接法展开)里已经给出说明了

又如:x→1是x-1→0,于是就有:

当x→1时,

√x=√[1+(x-1)]=1+(x-1)/2-(x-1)²/8+o((x-1)²)

总之,只要满足x→x₀时函数值□→0,那么我们就可以把复合函数√(1+□)按上述的泰勒公式在x=x₀处展开

那么不妨举一个"x→∞时函数值□→0"的例子看看:

当x→∞时,□=1/x→0,于是就有:

当x→∞时,

%5Cbegin%7Balign%7D%0A%5Csqrt%7B1%2B(%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20%20)%7D%26%3D1%2B%5Cfrac%7B1%7D%7B2%7D%20(%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20)-%5Cfrac%7B1%7D%7B8%7D%20(%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20)%5E2%2B%5Comicron%20((%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20%20)%5E2)%5C%5C%0A%26%3D1%2B%5Cfrac%7B1%7D%7B2%7D%5Cfrac%7B1%7D%7Bx%7D%20%20-%5Cfrac%7B1%7D%7B8%7D%20%5Cfrac%7B1%7D%7Bx%5E2%7D%20%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E2%7D%20)%0A%5Cend%7Balign%7D

可以预见,当外层展开的项数越多,余项在x→∞时就更快地收敛于0

ps:对于1/xᵖ(p>0)而言,p越大那么当x→∞时就更快地收敛于0(作商求比值的极限即可证明这个比阶的结论)

展开项数逐次递增的gif示意图如下:

所谓“无穷远处”的展开,也不过是复合函数泰勒展开的一种特殊情况罢了(特殊在自变量是趋于无穷内层函数□趋于x₀,且x₀又是外层函数可展开的点)

顺着这个思路进行拓展:

当x→+∞时,如果是√(1+x²)又该如何应对呢?

此时我们发现当x→+∞时x²→+∞,这时就不能再像

√(1+)=1+()/2-()²/8+o(()²)

这样展开了,毕竟这时x²是趋于+∞了,要用

√(1+)=1+□/2-□²/8+o(□²)

这个展开式就得要求内层函数值□→0才行


因此,这时就得“提取一个大头出来”:

√(1+x²)=x√(1/x²+1)

这时剩下的那个因式就可以用泰勒展开了(由于x→+∞时1/x²→0)

进而有:

x→+∞时,

%5Cbegin%7Balign%7D%0A%5Csqrt%7B1%2Bx%5E2%7D%26%3Dx%5Csqrt%7B%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D%20%2B1%20%7D%20%5C%5C%0A%26%3Dx%5B1%2B%5Cfrac%7B1%7D%7B2%7D(%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D)-%5Cfrac%7B1%7D%7B8%7D(%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D)%5E2%2B%5Comicron%20((%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D%20)%5E2)%20%20%5D%5C%5C%0A%26%3Dx%5B1%2B%5Cfrac%7B1%7D%7B2%7D%5Cfrac%7B1%7D%7Bx%5E2%7D%20-%5Cfrac%7B1%7D%7B8%7D%5Cfrac%7B1%7D%7Bx%5E4%7D%20%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E4%7D%20)%20%20%5D%5C%5C%0A%26%3Dx%2B%5Cfrac%7B1%7D%7B2%7D%5Cfrac%7B1%7D%7Bx%7D-%5Cfrac%7B1%7D%7B8%7D%20%20%5Cfrac%7B1%7D%7Bx%5E3%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E3%7D%20)%20%0A%5Cend%7Balign%7D


这里再简单提一下余项。当0<m≤n时,有:

x%5Cto%20%2B%5Cinfty%20%2Cx%5Em%20%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%3D%20%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E%7Bn-m%7D%7D%20)

这用高阶无穷小的定义即可证明:

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bx%20%5Cto%20%2B%5Cinfty%7D%20%5Cfrac%7Bx%5Em%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%7D%7B%5Cfrac%7B1%7D%7Bx%5E%7Bn-m%7D%7D%20%7D%3D%5Clim_%7Bx%20%5Cto%20%2B%5Cinfty%7D%20%5Cfrac%7B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%7D%7B%5Cfrac%7B1%7D%7Bx%5E%7Bn%7D%7D%20%7D%20%3D0%5C%5C%0A%5CRightarrow%20%26x%5Em%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%3D%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E%7Bn-m%7D%7D)%0A%5Cend%7Balign%7D

取m=1,n=4即得:x*o(1/x⁴)=o(1/x³),这也就是前面最后一步的由来

观察得到的展开式:

%5Csqrt%7B1%2Bx%5E2%7D%3Dx%2B%5Cfrac%7B1%7D%7B2%7D%5Cfrac%7B1%7D%7Bx%7D-%5Cfrac%7B1%7D%7B8%7D%20%20%5Cfrac%7B1%7D%7Bx%5E3%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E3%7D%20)%20

同样可以预见,展开项数越多,那么等式右边略去无穷小量后的曲线就越逼近前面的曲线,gif示意图如下:

当然了,因为是在无穷远处展开,因此这里的逼近当然就是指x→+∞时讨论的啦~


我们发现。有些函数具备如下特征:

当x→∞时,f(x)→∞,但f(x)仍可以用一个解析式形如p_m(x)%2Ba_1%5Cfrac%7B1%7D%7Bx%7D%2Ba_2%5Cfrac%7B1%7D%7Bx%5E2%7D%2B%5Ccdots%20%2Ba_n%20%5Cfrac%7B1%7D%7Bx%5En%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%20的曲线去逼近,即:

%5Cbegin%7Balign%7D%0Ax%5Cto%20%5Cinfty%20%2Cf(x)%26%3Dp_m(x)%2Ba_1%5Cfrac%7B1%7D%7Bx%7D%2Ba_2%5Cfrac%7B1%7D%7Bx%5E2%7D%2B%5Ccdots%20%2Ba_n%20%5Cfrac%7B1%7D%7Bx%5En%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%20%5C%5C%0A%26%3Dp_m(x)%2B%5Comicron%20(1)%0A%5Cend%7Balign%7D

其中pₘ(x)为一个m次多项式函数

ps:o(1)就表示无穷小量,这相当于次数为0的意思,即

lim o(1)=lim o(xº)/xº=0

我们就定义pₘ(x)为x→∞时函数f(x)的渐进m次曲线

特别地,当pₘ(x)为一次函数时,pₘ(x)就称为渐近线

为什么说是“渐进”呢?我们把pₘ(x)挪到左边就有:

%7B%5Csmall%20x%5Cto%20%5Cinfty%20%2Cf(x)-p_m(x)%3Da_1%5Cfrac%7B1%7D%7Bx%7D%2Ba_2%5Cfrac%7B1%7D%7Bx%5E2%7D%2B%5Ccdots%20%2Ba_n%20%5Cfrac%7B1%7D%7Bx%5En%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5En%7D%20)%3D%5Comicron(1)%20%5Cto%200%7D%20

这时右边剩下的项都是有限个x→∞时的无穷小量了,因而极限为0。这也就意味着x→∞时,f(x)与pₘ(x)的函数值之差可以任意小,因此就有“渐进”这么一说了

对于前面的这道题,就有:

√(1+x²)=x+o(1)

因此y=x就是函数f(x)=√(1+x²)x→+∞时的渐近线


另外,高中时期我们就学过双曲线x²/a²-y²/b²=1的渐近线是x²/a²-y²/b²=0,由于篇幅原因这里就留给读者当练习啦。方法跟这道题是一样的哦,换了个系数罢了~


例(2):求f(x)%3D%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%20%7D%5Csqrt%7B1%2Bx%5E2%7D在x→+∞时的渐近线

当x→+∞时,1/x→0,因此exp(1/x)可以展开;

而√(1+x²)前面已经讨论过了,需要先提一个“大头”出来再对剩下部分进行展开,即

f(x)%3Dx%5Ccdot%20%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%7D%20%5Csqrt%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%2B1%20%7D

这时就得应用准则(2)

%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%7D%20%5Csqrt%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%2B1%20%7D这个整体至少展开到-1阶(展开到含1/x的这一项),也即展成%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%7D%20%5Csqrt%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%2B1%20%7D%3Dc_0%2Bc_1%5Cfrac%7B1%7D%7Bx%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%7D)%20这种形式

这样才能保证x*o(1/x)=o(1)为无穷小量,也即保证x→+∞时余项极限为0

其中exp(1/x)的阶数为0(与xº=1同阶);

√(1/x²+1)极限为1,故阶数也为0(与xº同阶)

因此对于exp(1/x)而言,至少需要展开到0-1=-1阶;

对于√(1/x²+1)而言,至少需要展开到0-1=-1阶

于是

%5Cbegin%7Balign%7D%0A%26%5Cmathrm%7Be%7D%20%5E%7B%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20%7D%20%5Csqrt%7B%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D%20%2B1%20%7D%5C%5C%0A%3D%26%5B1%2B%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20%2B%5Comicron%20(%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20)%20%5D%5B1%2B%5Cfrac%7B1%7D%7B2%7D%20%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D%20%2B%5Comicron%20(%7B%5Ccolor%7BBlue%7D%20%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%7D%7D%20)%20%5D%5C%5C%0A%3D%26%5B1%2B%5Cfrac%7B1%7D%7Bx%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%7D%20)%20%5D%5B1%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%7D)%20%5D%5C%5C%0A%3D%26(1%2B%5Cfrac%7B1%7D%7Bx%7D%20)%5Ccdot%201%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%7D%20)%5C%5C%0A%3D%261%2B%5Cfrac%7B1%7D%7Bx%7D%20%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%7D%20)%0A%5Cend%7Balign%7D

进而有:

f(x)%3Dx%5Ccdot%20%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%7D%20%5Csqrt%7B%5Cfrac%7B1%7D%7Bx%5E2%7D%2B1%20%7D%3Dx%2B1%20%2B%5Comicron%20(1)

因此渐近线为y=x+1


例(3):求f(x)=√(x⁴-3x³+4)在x→+∞时的渐近曲线

同理,先需要先提一个“大头”出来再对剩下部分进行展开,即:

f(x)=x²√(1-3/x+4/x⁴)

当x→+∞时,-3/x+4/x⁴→0,进而可以用

□→0,√(1+)=1+□/2-□²/8+o(□²)

这个泰勒展开

由于大头是x²,因此要使得余项可以被忽略,就得让√(1-3/x+4/x⁴)至少展开到-2阶,而内层函数-3/x+4/x⁴~-3/x=O(1/x)为-1阶(与1/x同阶),因此外层至少要展开到第2项,即:

%5Cbegin%7Balign%7D%0A%26%5Csqrt%7B1%2B(%7B%5Ccolor%7BBlue%7D%7B%20-%5Cfrac%7B3%7D%7Bx%7D%2B%5Cfrac%7B4%7D%7Bx%5E4%7D%7D%7D%20%20)%20%20%7D%20%5C%5C%0A%3D%261%2B%5Cfrac%7B1%7D%7B2%7D(%7B%5Ccolor%7BBlue%7D%7B%20-%5Cfrac%7B3%7D%7Bx%7D%2B%5Cfrac%7B4%7D%7Bx%5E4%7D%7D%7D%20%20)%20-%5Cfrac%7B1%7D%7B8%7D(%7B%5Ccolor%7BBlue%7D%7B%20-%5Cfrac%7B3%7D%7Bx%7D%2B%5Cfrac%7B4%7D%7Bx%5E4%7D%7D%7D%20%20)%5E2%2B%5Comicron%20((%7B%5Ccolor%7BBlue%7D%7B%20-%5Cfrac%7B3%7D%7Bx%7D%2B%5Cfrac%7B4%7D%7Bx%5E4%7D%7D%7D%20)%5E2)%5C%5C%0A%3D%261-%5Cfrac%7B3%7D%7B2x%7D%20%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E2%7D%20)-%5Cfrac%7B1%7D%7B8%7D%20(%5Cfrac%7B9%7D%7Bx%5E2%7D%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E2%7D%20)%20)%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E2%7D%20)%5C%5C%0A%3D%261-%5Cfrac%7B3%7D%7B2%7D%5Cfrac%7B1%7D%7Bx%7D%20%20-%5Cfrac%7B9%7D%7B8%7D%20%5Cfrac%7B1%7D%7Bx%5E2%7D%20%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E2%7D%20)%0A%5Cend%7Balign%7D

进而有:

f(x)=x²√(1-3/x+4/x⁴)=x²-3x/2-9/8+o(1)

因此渐近曲线为:y=x²-3x/2-9/8

例(4):求极限%5Clim_%7Bx%20%5Cto%20%2B%5Cinfty%7D%20%5B(x%5E3-x%5E2%2B%5Cfrac%7Bx%7D%7B2%7D%20%2B1)%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%20%7D-%0A%5Csqrt%7Bx%5E6%2Bx%5E2%2Bx%2B1%7D%20%5D

ps:对于x→+∞还不太不熟悉的话,可以作个倒代换x=1/t化为t→0时的极限,然后在t=0处按4个展开准则进行展开即可。当然了,既然复合函数的泰勒展开也可以用于一些函数在无穷远处的展开(前面几道题皆是如此),那么这题也依旧可以用求渐进曲线的思路求解(毕竟极限存在就相当于渐近线是水平直线的这种再特殊些的情况呗)

两函数相减,由准则1,需要分别展开使得余项为o(1)

由于x³-x²+x/2+1大头为x³,因此要使得(x³-x²+x/2+1)exp(1/x)这个整体的余项为o(1),则需对exp(1/x)展开到0-3=-3阶(即展开到含1/x³这一项),于是就有:

%5Cbegin%7Balign%7D%0A%26(x%5E3-x%5E2%2B%5Cfrac%7Bx%7D%7B2%7D%20%2B1)%5Cmathrm%7Be%7D%20%5E%7B%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20%20%7D%5C%5C%0A%3D%26(x%5E3-x%5E2%2B%5Cfrac%7Bx%7D%7B2%7D%20%2B1)(1%2B(%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20)%2B%5Cfrac%7B(%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20)%5E2%7D%7B2!%7D%20%0A%2B%5Cfrac%7B(%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20%20)%5E3%7D%7B3!%7D%2B%5Comicron%20((%7B%5Ccolor%7BGreen%7D%20%7B%5Cfrac%7B1%7D%7Bx%7D%7D%7D%20)%5E3%20%20)%5C%5C%0A%3D%26(x%5E3-x%5E2%2B%5Cfrac%7Bx%7D%7B2%7D%20%2B1)(1%2B(%5Cfrac%7B1%7D%7Bx%7D)%2B%5Cfrac%7B(%5Cfrac%7B1%7D%7Bx%7D%20)%5E2%7D%7B2%7D%20%0A%2B%5Cfrac%7B(%5Cfrac%7B1%7D%7Bx%7D%20)%5E3%7D%7B6%7D)%2B%5Comicron%20(1)%5C%5C%0A%3D%26x%5E3%2B%5Cfrac%7B7%7D%7B6%7D%20%2B%5Comicron%20(1)%0A%5Cend%7Balign%7D


再展开另一项:

√(x⁶+x²+x+1)=x³√[1+(1/x⁴+1/x⁵+1/x⁶)]

前面练习过好几次这种形式的展开啦,都是先提取大头再对剩下部分进行展开

x→+∞时,1/x⁴+1/x⁵+1/x⁶→0

且由于大头为x³,因此要使得乘开后的余项为o(1),

√[1+(1/x⁴+1/x⁵+1/x⁶)]至少需要展开到-3阶(即展开到含1/x³这一项)

又由于内层函数1/x⁴+1/x⁵+1/x⁶~1/x⁴,阶数为-4,因此外层至少展开到第2项,即

%5Cbegin%7Balign%7D%0A%26%5Csqrt%7B1%2B%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E4%7D%2B%5Cfrac%7B1%7D%7Bx%5E5%7D%2B%5Cfrac%7B1%7D%7Bx%5E6%7D%20%7D%7D%20%20%20%7D%20%5C%5C%0A%3D%261%2B%5Cfrac%7B1%7D%7B2%7D%20(%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E4%7D%2B%5Cfrac%7B1%7D%7Bx%5E5%7D%2B%5Cfrac%7B1%7D%7Bx%5E6%7D%20%7D%7D%20%20)%0A%2B%5Comicron%20(%7B%5Ccolor%7BBlue%7D%7B%20%5Cfrac%7B1%7D%7Bx%5E4%7D%2B%5Cfrac%7B1%7D%7Bx%5E5%7D%2B%5Cfrac%7B1%7D%7Bx%5E6%7D%20%7D%7D%20%20)%5C%5C%0A%3D%261%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E3%7D%20)%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E4%7D%20)%5C%5C%0A%3D%261%2B%5Comicron%20(%5Cfrac%7B1%7D%7Bx%5E3%7D%20)%0A%5Cend%7Balign%7D

进而有:

√(x⁶+x²+x+1)=x³√[1+(1/x⁴+1/x⁵+1/x⁶)]=x³+o(1)


综上得:

%5Cbegin%7Balign%7D%0A%26(x%5E3-x%5E2%2B%5Cfrac%7Bx%7D%7B2%7D%20%2B1)%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B1%7D%7Bx%7D%20%7D-%0A%5Csqrt%7Bx%5E6%2Bx%5E2%2Bx%2B1%7D%5C%5C%0A%3D%26x%5E3%2B%5Cfrac%7B7%7D%7B6%7D%20%2B%5Comicron%20(1)-(x%5E3%2B%5Comicron%20(1))%5C%5C%0A%3D%26%5Cfrac%7B7%7D%7B6%7D%20%2B%5Comicron%20(1)%0A%5Cend%7Balign%7D

故原极限为7/6


补充完了渐进曲线的知识,下面再补充几个常用的泰勒展开

补充几个常用的泰勒展开

%5Cbbox%5B%238fffa8%5D%7B%5Cbegin%7Balign%7D%0A%5Cfrac%7B%5Cmathrm%7Be%7D%20%5Ex-%5Cmathrm%7Be%7D%20%5E%7B-x%7D%7D%7B2%7D%26%20%3D%5Csinh%20x%3Dx%2B%5Cfrac%7Bx%5E3%7D%7B3!%7D%2B%5Cfrac%7Bx%5E5%7D%7B5!%7D%2B%5Ccdots%20%2B%5Cfrac%7Bx%5E%7B2n%2B1%7D%7D%7B(2n%2B1)!%7D%2B%5Comicron%20(x%5E%7B2n%2B1%7D)%20%5C%5C%0A%5Cfrac%7B%5Cmathrm%7Be%7D%20%5Ex%2B%5Cmathrm%7Be%7D%20%5E%7B-x%7D%7D%7B2%7D%20%26%3D%5Ccosh%20x%3D1%2B%5Cfrac%7Bx%5E2%7D%7B2!%7D%2B%5Cfrac%7Bx%5E4%7D%7B4!%7D%2B%5Ccdots%20%2B%5Cfrac%7Bx%5E%7B2n%7D%7D%7B(2n)!%7D%2B%5Comicron%20(x%5E%7B2n%7D)%20%5C%5C%0A%5Ctan%20x%26%3Dx%2B%5Cfrac%7Bx%5E3%7D%7B3%7D%2B%5Cfrac%7B2%7D%7B15%7Dx%5E5%2B%5Comicron%20(x%5E5)%20%20%5C%5C%0A%5Cfrac%7B(1%2Bx)%5E%7B%5Cfrac%7B1%7D%7Bx%7D%20%7D%7D%7B%5Cmathrm%7Be%7D%20%7D%20%26%3D1-%5Cfrac%7B1%7D%7B2%7Dx%2B%5Cfrac%7B11%7D%7B24%7Dx%5E2-%5Cfrac%7B7%7D%7B16%7Dx%5E3%20%20%2B%5Comicron%20(x%5E3)%20%20%0A%5Cend%7Balign%7D%7D

以上几个函数均在x=0处展开

前两者分别是双曲正弦/双曲余弦函数,其与三角函数是孪生兄弟,因此展开式也很好类比,这里的sinhx,coshx的展开式就没有正负交错,除掉正负号后系数与sinx,cosx的完全一致。

这两个可以用直接法推(直接求n阶导归纳),也可以用间接法推:由eᕽ的展开式,推出e⁻ᕽ展开式,进而推得sinhx=(eᕽ-e⁻ᕽ)/2和coshx=(eᕽ+e⁻ᕽ)/2的展开式

tanx是直接淦前5阶导的,虽然也有通项公式,但涉及伯努利数以及较长的证明过程,这里就先略去推导了。

最后一个幂指函数是通过取指数化为如下的复合函数:

%5Cfrac%7B(1%2Bx)%5E%7B%5Cfrac%7B1%7D%7Bx%7D%20%7D%7D%7B%5Cmathrm%7Be%7D%20%7D%20%3D%5Cmathrm%7Bexp%7D%20%5B%5Cfrac%7B%5Cln%20(1%2Bx)-x%7D%7Bx%7D%5D

然后用复合函数的泰勒展开去推得

最后这个展开式有空位就推(写完发现公式数满了,就留作感兴趣的读者温故而知新吧~)

练习题

接下来就是极限的练习题啦,读者们可通过下面的练习巩固4个选阶准则哦~

题(1):求极限%5Clim_%7Bx%20%5Cto%200%7D%20(%5Cfrac%7B%5Carcsin%20x%7D%7B%5Csin%20x%7D%20)%5E%7B%5Ccot%5E2%20x%7D

运用的准则:准则(1)

以下证明更一般的情况的处理方法。对于1^∞型幂指函数极限,可以按下面的“三部曲”转化为∞*0的未定式

已知lim f=1,lim g=∞,则有:

%5Cbegin%7Balign%7D%0A%26%5Clim%20%20f%5Eg%5C%5C%0A%3D%26%5Clim%20%20%5Cmathrm%7Bexp%7D%20%5B%5Cln%20(f%5Eg)%5D%5C%5C%0A%3D%26%5Clim%20%20%5Cmathrm%7Bexp%7D%20(g%5Cln%20f)%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20(%5Clim%20%20%20g%5Cln%20f)%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim%20%20%20g(f-1)%5D%0A%5Cend%7Balign%7D

前面两行是恒等变形,幂指函数都可以利用这种方法化为一个复合函数

第4行是利用了外层的指数函数的连续性,由第二复合法则,可交换取极限和复合的次序,也就是将极限符号移至指数里

最后一步利用了等价无穷小:□→1,ln□~□-1

回到本题。经判断,这题是1^∞未定式,由上述步骤立刻可化为:

%5Cbegin%7Balign%7D%0A%5Clim_%7Bx%20%5Cto%200%7D%20(%5Cfrac%7B%5Carcsin%20x%7D%7B%5Csin%20x%7D%20)%5E%7B%5Ccot%5E2%20x%7D%26%3D%0A%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bx%20%5Cto%200%7D%5Ccot%5E2x%20(%5Cfrac%7B%5Carcsin%20x%7D%7B%5Csin%20x%7D-1)%5D%5C%5C%0A%26%3D%5Cmathrm%7Bexp%7D%20(%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Carcsin%20x-%5Csin%20x%7D%7B%5Csin%20x%5Ctan%5E2%20x%7D)%0A%5Cend%7Balign%7D

其中分母均为因式,用等价即可:

x→0,sinx~x,tanx~x,进而sinx tan²x~x³,可知分母为3阶

ps:对于可展开的函数(某处n阶可导)而言,等价无穷小相当于展开式取到1阶的情形

有了分母阶数的提示,需将分子也展开到3阶,进而由准则(1),需将arcsinx,sinx分别展开到3阶,即有:

%5Cbegin%7Balign%7D%0A%26%5Cmathrm%7Bexp%7D%20(%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Carcsin%20x-%5Csin%20x%7D%7B%5Csin%20x%5Ctan%5E2%20x%7D)%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7Bx%2B%5Cfrac%7B1!!%7D%7B2!!%5Ccdot%203%7D%20x%5E3%2B%5Comicron%20(x%5E3)-(x-%5Cfrac%7Bx%5E3%7D%7B3!%7D%2B%5Comicron%20(x%5E3)%20)%7D%7Bx%5Ccdot%20x%5E2%7D%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Cfrac%7Bx%5E3%7D%7B3%7D%20%2B%5Comicron%20(x%5E3)%20%7D%7Bx%5E3%7D%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20(%5Cfrac%7B1%7D%7B3%7D%20)%5C%5C%0A%3D%26%5Csqrt%5B3%5D%7B%5Cmathrm%7Be%7D%20%7D%20%0A%5Cend%7Balign%7D


题(2):求极限%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Ccos%20x-%5Ccos%20(%5Csin%20x)%7D%7Bx%5E4%7D%20

运用的准则:准则(1),准则(2),准则(3)

由于分母是4次,因此分子也得展开到4次,由准则(1),需将cosx,cos(sinx)分别展开到4次

其中cosx=1-x²/2!+x⁴/4!+o(x⁴)

而cos(sinx)是复合函数,由sinx~x,因此外层需要展开到4阶,即:

cos(sinx)=1-(sinx)²/2!+(sinx)⁴/4!+o((sinx))

由于sinx~x,故o(sin⁴x)=o(x⁴),接下来将前面的sin²x,sin⁴x也展开到4阶,这时是函数乘积,因此就得用准则(2)

sinx阶数为1(与x同阶)

sin²x=sinx*sinx,对于第一个sinx,需要展开的阶数为:4-1=3

同理对另一个sinx,也需展开到4-1=3阶

sin²x=[x-x³/3!+o(x³)]²

=(x-x³/3!)²+o(x⁴)

=x²-x⁴/3+o(x⁴)

sin⁴x=sinx*sinx*sinx*sinx,对于第一个sinx,需要展开的阶数为:4-(1+1+1)=1

同理对其余几个sinx,也需展开到4-(1+1+1)=1阶

sin⁴x=[x+o(x)]⁴=x⁴+o(x⁴)

综上得:

%5Cbegin%7Balign%7D%0A%5Ccos%20(%5Csin%20x)%26%3D1-%5Cfrac%7B(%5Csin%20x)%5E2%7D%7B2!%7D%2B%5Cfrac%7B(%5Csin%20x)%5E4%7D%7B4!%7D%2B%5Comicron%20%20(%5Csin%5E4x)%20%20%5C%5C%0A%26%3D1-%5Cfrac%7Bx%5E2-%5Cfrac%7Bx%5E4%7D%7B3%7D%2B%5Comicron%20%20(x%5E4)%7D%7B2%7D%2B%5Cfrac%7Bx%5E4%2B%5Comicron%20%20(x%5E4)%7D%7B24%7D%2B%5Comicron%20%20(x%5E4)%20%20%5C%5C%0A%26%3D1-%5Cfrac%7Bx%5E2%7D%7B2%7D%20%2B%5Cfrac%7B5%7D%7B24%7D%20x%5E4%2B%5Comicron%20%20(x%5E4)%20%20%0A%5Cend%7Balign%7D

最终得:

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Ccos%20x-%5Ccos%20(%5Csin%20x)%7D%7Bx%5E4%7D%20%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B1-%5Cfrac%7Bx%5E2%7D%7B2%7D%2B%5Cfrac%7Bx%5E4%7D%7B24%7D%2B%5Comicron%20%20(x%5E4)%0A-(1-%5Cfrac%7Bx%5E2%7D%7B2%7D%20%2B%5Cfrac%7B5%7D%7B24%7D%20x%5E4%2B%5Comicron%20%20(x%5E4)%20%20)%20%20%7D%7Bx%5E4%7D%20%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B-%5Cfrac%7Bx%5E4%7D%7B6%7D%2B%5Comicron%20(x%5E4)%20%7D%7Bx%5E4%7D%20%5C%5C%0A%3D%26-%5Cfrac%7B1%7D%7B6%7D%20%0A%5Cend%7Balign%7D


以上是泰勒展开的解法,对于涉及复合函数的泰勒展开,就需要使用准则(3),使用完准则(3)后,剩下的部分又是若干个内层函数次幂的形式,这时就得再用准则(2)逐个展开,可见对复合函数泰勒展开时,在计算这一块上一定要打起十二分精神哦~

当然了,就这题而言还有更简单些的解法:

注意到分子是cos()-cos()的形式,因此可以考虑使用和差化积分解成两个因式:

cosx-cos(sinx)=-2sin[(x+sinx)/2] sin[(x-sinx)/2]

这样处理会带来什么好处呢?此时右边已经分解成两个因子了,因此我们求极限的时候就可以分而治之,分别对这两个因子进行估阶即可

这时可能又有新的疑惑,你这个虽然是拆成了两个因子,但这两个都还是复合函数,况且两函数相乘还得用准则(2)呢,哪里简化了?

%5Clim%20%20%5Cfrac%7Bf_1%20%5Ccdot%20f_2%5Ccdots%20f_m%7D%7Bg_1%5Ccdot%20g_2%5Ccdots%20g_n%7D%20(注意这时候f₁,f₂,...,fₘ,g₁,g₂,...,gₙ对于解析式整体而言因子)

两个关键词,一个是“整体”,一个是“因子”,相信大家学到等价无穷小的时候一定有特别注意这点

如果因子的极限非0,那只需直接代入即可;

如果因子的极限为0,那只需找其等价无穷小(对于可泰勒展开的函数,就相当于只用展开到x¹项,这样对于外层函数展开的工作量就大大减小了),从而就达成了分而治之的目的。

这两者都是极限乘法法则保证的,非零因子代入、零因子等价的实际步骤(例子如下):

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7Bf(x)%7D%7B%5Ccos%20x%7D%20%3D%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B1%7D%7B%5Ccos%20x%7D%5Ccdot%20%5Clim_%7Bx%20%5Cto%200%7D%20f(x)%3D%5Cfrac%7B1%7D%7B%5Ccos%200%7D%5Ccdot%20%20%5Clim_%7Bx%20%5Cto%200%7D%20f(x)%5C%5C%0A%26%5C%5C%0A%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7Bf(x)%7D%7B%5Csin%20%20x%7D%20%3D%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7Bx%7D%7B%5Csin%20x%7D%5Ccdot%20%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7Bf(x)%7D%7Bx%7D%3D1%5Ccdot%20%20%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7Bf(x)%7D%7Bx%7D%0A%5Cend%7Balign%7D


回到原题,即有:

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Ccos%20x-%5Ccos%20(%5Csin%20x)%7D%7Bx%5E4%7D%20%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B-2%5Csin%20(%5Cfrac%7Bx%2B%5Csin%20x%7D%7B2%7D%20)%5Csin%20(%5Cfrac%7Bx-%5Csin%20x%7D%7B2%7D%20)%7D%7Bx%5E4%7D%5C%5C%0A%3D%26%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B-2%5Ccdot%20%5Cfrac%7Bx%2B%5Csin%20x%7D%7B2%7D%20%5Ccdot%20%5Cfrac%7Bx-%5Csin%20x%7D%7B2%7D%20%7D%7Bx%5E4%7D%5C%5C%0A%3D%26-%5Cfrac%7B1%7D%7B2%7D%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Bx%2Bx%2B%5Comicron%20(x)%5D%5Bx-(x-%5Cfrac%7Bx%5E3%7D%7B3!%7D%2B%5Comicron%20(x%5E3)%20)%5D%7D%7Bx%5E4%7D%5C%5C%0A%3D%26-%5Cfrac%7B1%7D%7B2%7D%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B2x%5Ccdot%20%5Cfrac%7B1%7D%7B6%7Dx%5E3%20%7D%7Bx%5E4%7D%5C%5C%0A%3D%26-%5Cfrac%7B1%7D%7B6%7D%20%0A%5Cend%7Balign%7D


此外,本题还可以使用拉格朗日中值定理,由于篇幅有限,这里就先不展开介绍了

对于像本题一样可以通过适当的代数变形完成因式分解的主要是如下几种形式:

%5Cbbox%5B%238fffa8%5D%7B%5Cbegin%7Balign%7D%0A%26%5Csin%20%5Calpha%20%2B%5Csin%20%5Cbeta%20%3D2%5Csin%20%5Cfrac%7B%5Calpha%20%2B%5Cbeta%7D%7B2%7D%5Ccos%20%5Cfrac%7B%5Calpha%20-%5Cbeta%7D%7B2%7D%5C%5C%0A%26%5Csin%20%5Calpha%20-%5Csin%20%5Cbeta%20%3D2%5Ccos%20%5Cfrac%7B%5Calpha%20%2B%5Cbeta%7D%7B2%7D%5Csin%20%5Cfrac%7B%5Calpha%20-%5Cbeta%7D%7B2%7D%5C%5C%0A%26%5Ccos%20%5Calpha%20%2B%5Ccos%20%5Cbeta%20%3D2%5Ccos%20%5Cfrac%7B%5Calpha%20%2B%5Cbeta%7D%7B2%7D%5Ccos%20%5Cfrac%7B%5Calpha%20-%5Cbeta%7D%7B2%7D%5C%5C%0A%26%5Ccos%20%5Calpha%20-%5Ccos%20%5Cbeta%20%3D-2%5Csin%20%5Cfrac%7B%5Calpha%20%2B%5Cbeta%7D%7B2%7D%5Csin%20%5Cfrac%7B%5Calpha%20-%5Cbeta%7D%7B2%7D%5C%5C%0A%26%5Cln%20a%2B%5Cln%20b%3D%5Cln%20(ab)%2C%5Cln%20a-%5Cln%20b%3D%5Cln%20(%5Cfrac%7Ba%7D%7Bb%7D%20)%5C%5C%0A%26%5Cmathrm%7Be%7D%20%5Ea-%5Cmathrm%7Be%7D%20%5Eb%3D%5Cmathrm%7Be%7D%20%5Eb(%5Cmathrm%7Be%7D%20%5E%7Ba-b%7D-1)%0A%5Cend%7Balign%7D%7D

前4个是和差化积公式,三角函数的基本公式要记住的哦

推导也不难,得用和差角公式+换元,可自行查找相关资料

后面两个分别是对数运算/指数运算

最后一个为什么要提一个eᵃeᵇ出来呢?这是由于在求极限时,如果我们遇到exp(f)-exp(g)这种因式,且f和g极限存在且相同,这时处理成右边这种形式后,eᵃ是非零因子直接代入,eᵃ⁻ᵇ-1是零因子,可以用等价:a-b→0,eᵃ⁻ᵇ-1~a-b

下面几道变式训练题,利用优先因式分解的思路会比上来就泰勒来得快:

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bx%20%5Cto%200%7D%20%5B%5Cfrac%7B1%7D%7B%5Cln%20(1%2Bx%5E2)%7D-%5Cfrac%7B1%7D%7B%5Cln%20(1%2B%5Ctan%20%5E2x)%7D%20%20%5D%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Cln%20(1%2B%5Ctan%5E2%20x)-%5Cln%20(1%2Bx%5E2)%7D%7B%5Cln%20(1%2Bx%5E2)%5Cln%20(1%2B%5Ctan%5E2%20x)%7D%5C%5C%0A%3D%26%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%7B%5Ccolor%7BBlue%7D%20%7B%5Cln%20(%5Cfrac%7B1%2B%5Ctan%5E2%20x%7D%7B1%2Bx%5E2%7D%20)%7D%7D%20%7D%7Bx%5E2%5Ccdot%20%5Ctan%5E2%20x%7D%5C%5C%0A%3D%26%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Cfrac%7B1%2B%5Ctan%5E2%20x%7D%7B1%2Bx%5E2%7D-1%7D%7Bx%5E4%7D%5C%5C%0A%3D%26%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Ctan%5E2%20x-x%5E2%7D%7Bx%5E4(1%2Bx%5E2)%7D%5C%5C%0A%3D%26%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%7B%5Ccolor%7BBlue%7D%20%7B(%5Ctan%20x%2Bx)(%5Ctan%20x-x)%7D%7D%20%7D%7Bx%5E4%7D%5C%5C%0A%3D%26%20%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B%5Bx%2B%5Comicron%20(x)%2Bx%5D%5Bx%2B%5Cfrac%7Bx%5E3%7D%7B3%7D%2B%5Comicron%20(x%5E3)%20-x%5D%7D%7Bx%5E4%7D%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%5Cfrac%7B2x%5Ccdot%20%5Cfrac%7Bx%5E3%7D%7B3%7D%7D%7Bx%5E4%7D%5C%5C%0A%3D%26%5Cfrac%7B2%7D%7B3%7D%20%0A%5Cend%7Balign%7D


%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Ccos%20(x%20%5Cmathrm%7Be%7D%20%5Ex)-%5Ccos%20(x%20%5Cmathrm%7Be%7D%20%5E%7B-x%7D)%7D%7Bx%5E3%7D%20%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%7B%5Ccolor%7BBlue%7D%20%7B-2%5Csin%20(x%5Ccdot%20%5Cfrac%7B%5Cmathrm%7Be%7D%5Ex%2B%5Cmathrm%7Be%7D%5E%7B-x%7D%7D%7B2%7D)%20%5Csin%20(x%20%5Ccdot%20%5Cfrac%7B%5Cmathrm%7Be%7D%5Ex-%5Cmathrm%7Be%7D%5E%7B-x%7D%7D%7B2%7D)%7D%7D%20%7D%7Bx%5E3%7D%20%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B-2%5Ccdot%20(x%5Ccdot%20%5Cfrac%7B%5Cmathrm%7Be%7D%5Ex%2B%5Cmathrm%7Be%7D%5E%7B-x%7D%7D%7B2%7D)%0A%5Ccdot%20(x%20%5Ccdot%20%5Cfrac%7B%5Cmathrm%7Be%7D%5Ex-%5Cmathrm%7Be%7D%5E%7B-x%7D%7D%7B2%7D)%7D%7Bx%5E3%7D%20%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B-2%5Ccdot%20x%5Ccdot%201%0A%5Ccdot%20x%20%5Ccdot%20x%7D%7Bx%5E3%7D%20%5C%5C%0A%3D%26-2%0A%5Cend%7Balign%7D

倒数第二步coshx极限为1,非零因子可以代入。

sinhx为零因子,可用等价x→0,sinhx~x,这个用补充内容里的泰勒展开可推哦

其中标蓝的部分就是用到了因式分解思路

由此可见,泰勒展开是最普适的方法,但对于某些题而言计算量会稍大。因此得讲求极限的方法融会贯通才能尽最大可能选出最合适的方法~

题(3):求极限%5Clim_%7Bx%20%5Cto%200%7D%20(-%5Cfrac%7B%5Ccot%20x%7D%7B%5Cmathrm%7Be%7D%20%5E%7B2x%7D%7D%20%2B%5Cfrac%7B1%7D%7B%5Cmathrm%7Be%7D%20%5E%7B-x%7D%5Csin%5E2x%20%7D-%5Cfrac%7B1%7D%7Bx%5E2%7D%20%20)

运用的准则:准则(1),准则(2),准则(3)

先对函数进行通分化简:

%5Cbegin%7Balign%7D%0A%26-%5Cfrac%7B%5Ccot%20x%7D%7B%5Cmathrm%7Be%7D%20%5E%7B2x%7D%7D%20%2B%5Cfrac%7B1%7D%7B%5Cmathrm%7Be%7D%20%5E%7B-x%7D%5Csin%5E2x%20%7D-%5Cfrac%7B1%7D%7Bx%5E2%7D%5C%5C%0A%3D%26-%5Cfrac%7B%5Ccos%20x%7D%7B%5Cmathrm%7Be%7D%20%5E%7B2x%7D%5Csin%20x%7D%2B%5Cfrac%7B%5Cmathrm%7Be%7D%20%5Ex%7D%7B%5Csin%5E2%20x%7D-%5Cfrac%7B1%7D%7Bx%5E2%7D%20%5C%5C%0A%3D%26%5Cfrac%7B-x%5E2%5Csin%20x%5Ccos%20x%2Bx%5E3%5Cmathrm%7Be%7D%20%5E%7B3x%7D-%5Cmathrm%7Be%7D%20%5E%7B2x%7D%5Csin%5E2%20x%7D%7Bx%5E2%5Cmathrm%7Be%7D%20%5E%7B2x%7D%5Csin%5E2%20x%7D%5C%5C%0A%3D%26%5Cfrac%7B-x%5E2%5Csin%202x%2B2x%5E2%20%5Cmathrm%7Be%7D%20%5E%7B3x%7D%2B%5Cmathrm%7Be%7D%20%5E%7B2x%7D(%5Ccos%202x-1)%7D%7B2x%5E2%5Cmathrm%7Be%7D%20%5E%7B2x%7D%5Csin%5E2%20x%7D%20%20%20%20%0A%5Cend%7Balign%7D

这里用到的切化弦,这是由于sinx和cosx的展开式比tanx的好推导且方便记忆。

最后一步用了二倍角公式处理,简化之处在于将两个函数相乘sinxcosx,-sin²x的形式分别和并成了一个函数sin2x,cos2x-1,这样稍加处理也便于后续展开

分母全为因子,考虑先对分母估阶:

x→0,e²ˣ→1,sinx~x,因此分母2x²e²ˣsin²x~2x⁴,阶数为4

有了分母阶数的提示,那么分子至少得展开到4阶

准则(1),需要将-x²sin2x,2x²e³ˣ,e²ˣ(cos2x-1)分别展开到4阶


接下来由准则(2)分别对这3者进行展开:

对于-x²sin2x,由于x²阶数为2,故sin2x需展开到4-2=2阶,

而2x阶数为1(与x同阶),因此需要将外层的sin展开到2阶,即有:

-x²sin2x=-x²[2x+o((2x)²)]=-2x³+o(x⁴)


对于2x²e³ˣ同理,由于x²阶数为2,故e³ˣ需展开到4-2=2阶,

而2x阶数为1(与x同阶),因此需要将外层的指数函数展开到2阶,即有:

2x²e³ˣ=2x²[1+(3x)+(3x)²/2!+o((3x)²)]

=2x²(1+3x+9x²/2)+o(x⁴)

=2x²+6x³+9x⁴+o(x⁴)

对于e²ˣ(cos2x-1),分别求出两个函数的阶数:

e²ˣ阶数为0(与xº=1同阶);

cos2x-1=1-(2x)²/2!+o((2x)²)-1=-2x²+o(x²)~-2x²=O(x²),阶数为2(与x²同阶)

因此由准则(2)

对于e²ˣ,需要展开到4-2=2阶;

对于cos2x-1,需展开到4-0=4阶

即有:

%5Cbegin%7Balign%7D%0A%26%5Cmathrm%7Be%7D%20%5E%7B%7B%5Ccolor%7BGreen%7D%20%7B2x%7D%7D%20%7D(%5Ccos%20%7B%5Ccolor%7BBlue%7D%20%7B2x%7D%7D%20-1)%5C%5C%0A%3D%26%5B1%2B%7B%5Ccolor%7BGreen%7D%20%7B2x%7D%7D%20%2B%5Cfrac%7B(%7B%5Ccolor%7BGreen%7D%20%7B2x%7D%7D%20)%5E2%7D%7B2!%7D%20%2B%5Comicron%20((%7B%5Ccolor%7BGreen%7D%20%7B2x%7D%7D%20)%5E2)%5D%5B1-%5Cfrac%7B(%7B%5Ccolor%7BBlue%7D%20%7B2x%7D%7D)%5E2%7D%7B2!%7D%2B%5Cfrac%7B(%7B%5Ccolor%7BBlue%7D%20%7B2x%7D%7D)%5E4%7D%7B4!%7D%20%2B%5Comicron%20((%7B%5Ccolor%7BBlue%7D%20%7B2x%7D%7D)%5E4)%20-1%5D%5C%5C%0A%3D%26%5B1%2B2x%2B2x%5E2%20%2B%5Comicron%20(x%5E2)%5D%5B-2x%5E2%2B%5Cfrac%7B2%7D%7B3%7D%20x%5E4%20%2B%5Comicron%20(x%5E4)%20%5D%5C%5C%0A%3D%26(1%2B2x%2B2x%5E2%20)(-2x%5E2%2B%5Cfrac%7B2%7D%7B3%7D%20x%5E4%20)%2B%5Comicron%20(x%5E4)%5C%5C%0A%3D%26-2x%5E2-4x%5E3-%5Cfrac%7B10%7D%7B3%7Dx%5E4%20%2B%5Comicron%20(x%5E4)%0A%5Cend%7Balign%7D

综上得:

%5Cbegin%7Balign%7D%0A%5Ctext%7B%E5%8E%9F%E5%BC%8F%7D%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B-x%5E2%5Csin%202x%2B2x%5E2%20%5Cmathrm%7Be%7D%20%5E%7B3x%7D%2B%5Cmathrm%7Be%7D%20%5E%7B2x%7D(%5Ccos%202x-1)%7D%7B2x%5E2%5Cmathrm%7Be%7D%20%5E%7B2x%7D%5Csin%5E2%20x%7D%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5B-2x%5E3%2B%5Comicron%20(x%5E4)%5D%2B%5B2x%5E2%2B6x%5E3%2B9x%5E4%2B%5Comicron%20(x%5E4)%5D%2B%5B-2x%5E2-4x%5E3-%5Cfrac%7B10%7D%7B3%7Dx%5E4%20%2B%5Comicron%20(x%5E4)%5D%7D%7B2x%5E4%7D%5C%5C%0A%3D%26%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cfrac%7B17%7D%7B3%7Dx%5E4%20%2B%5Comicron%20(x%5E4)%7D%7B2x%5E4%7D%5C%5C%0A%3D%26%5Cfrac%7B17%7D%7B6%7D%0A%5Cend%7Balign%7D


题(4):求极限%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cint_%7B0%7D%5E%7Bx%7Dt%5Csin%20t%5Cmathrm%7Bd%7Dt%20%20%7D%7B%5Csin%20x-%5Ctan%20x%7D%20

由于分子分母各自都只有1个因式,因此只需分别找其等价无穷小即可

分母直接用最基础的准则(1),分别展开直到首个系数不同的项即可:

sinx=x-x³/6+o(x³)

tanx=x+x³/3+o(x³)

⇒sinx-tanx=-x³/2+o(x³)~-x³/2

下面再处理分子,对于分子这种变限积分形式的等价,以下给出一个更一般的结论:

x%5Cto%200%2Cf(x)%5Csim%20a%20x%5E%5Calpha%20%2Cg(x)%5Csim%20b%20x%5E%5Cbeta,则有:

%5Clim_%7Bx%20%5Cto%200%7D%20%5Cint_%7B0%7D%5E%7Bg(x)%7D%20f(t)%5Cmathrm%7Bd%7Dt%20%5Csim%20%5Cint_%7B0%7D%5E%7Bb%20x%5E%5Cbeta%20%7D%20at%5E%5Calpha%20%5Cmathrm%7Bd%7Dt

注意这里的下限是0哦(下限也是函数的话就不能直接等价了)。此时可将被积函数积分上限替换为对应的等价无穷小,积分出来的结果就是原式的等价无穷小

证明:

由已知得:f(x)=a x^α+o(x^α)

准则(4)可得:

%5Cbegin%7Balign%7D%0A%26%5Cint_%7B0%7D%5E%7Bx%7D%20f(t)%5Cmathrm%7Bd%7Dt%20%5C%5C%0A%3D%26%5Cint_%7B0%7D%5E%7Bx%7D%20ax%5E%5Calpha%20%5Cmathrm%7Bd%7Dt%20%2B%5Comicron%20(x%5E%7B%5Calpha%20%2B1%7D)%5C%5C%0A%3D%26%5Cfrac%7Ba%7D%7B%5Calpha%20%2B1%7Dx%5E%7B%5Calpha%20%2B1%7D%2B%5Comicron%20(x%5E%7B%5Calpha%20%2B1%7D)%5C%5C%0A%5Csim%26%20%20%5Cfrac%7Ba%7D%7B%5Calpha%20%2B1%7Dx%5E%7B%5Calpha%20%2B1%7D%0A%5Cend%7Balign%7D

再由准则(3)得:

%5Cbegin%7Balign%7D%0A%26%5Cint_%7B0%7D%5E%7Bb%20x%5E%5Cbeta%20%7D%20f(t)%5Cmathrm%7Bd%7Dt%20%5C%5C%0A%3D%26%5Cint_%7B0%7D%5E%7Bbx%5E%5Cbeta%20%7D%20ax%5E%5Calpha%20%5Cmathrm%7Bd%7Dt%20%2B%5Comicron%20((bx%5E%5Cbeta%20)%5E%7B%5Calpha%20%2B1%7D)%5C%5C%0A%3D%26%5Cfrac%7Ba%7D%7B%5Calpha%20%2B1%7D(bx%5E%5Cbeta%20)%5E%7B%5Calpha%20%2B1%7D%2B%5Comicron%20((bx%5E%5Cbeta%20)%5E%7B%5Calpha%20%2B1%7D)%5C%5C%0A%5Csim%26%20%20%5Cfrac%7Ba%7D%7B%5Calpha%20%2B1%7D(bx%5E%5Cbeta%20)%5E%7B%5Calpha%20%2B1%7D%0A%5Cend%7Balign%7D

结论得证!

用到的其实就是上一篇文章 泰勒公式间接展开法(拔高篇) 中的定理3定理2

结论的应用例子:

%5Cbegin%7Balign%7D%0A%26x%5Cto%200%2C%5C%5C%0A%26%5Cint_%7B0%7D%5E%7Bx%7D%20(%5Cmathrm%7Be%7D%20%5E%7Bt%5E2%7D-1)%5Cmathrm%7Bd%7Dt%20%5Csim%20%5Cint_%7B0%7D%5E%7Bx%7D%20t%5E2%5Cmathrm%7Bd%7Dt%3D%5Cfrac%7B1%7D%7B3%7D%20x%5E3%5C%5C%0A%26%5Cint_%7B0%7D%5E%7B1-%5Ccos%20x%7D%5Csin%20(t%5E2)%5Cmathrm%7Bd%7Dt%20%5Csim%20%20%20%5Cint_%7B0%7D%5E%7B%5Cfrac%7Bx%5E2%7D%7B2%7D%20%7Dt%5E2%5Cmathrm%7Bd%7Dt%3D%5Cfrac%7Bx%5E6%7D%7B24%7D%20%0A%5Cend%7Balign%7D


由这个结论,立马就有:

%5Cint_%7B0%7D%5E%7Bx%7Dt%5Csin%20t%5Cmathrm%7Bd%7Dt%20%5Csim%20%5Cint_%7B0%7D%5E%7Bx%7Dt%5E2%5Cmathrm%7Bd%7Dt%3D%5Cfrac%7B1%7D%7B3%7D%20x%5E3

进而

%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cint_%7B0%7D%5E%7Bx%7Dt%5Csin%20t%5Cmathrm%7Bd%7Dt%20%20%7D%7B%5Csin%20x-%5Ctan%20x%7D%20%0A%3D%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cfrac%7B1%7D%7B3%7Dx%5E3%20%7D%7B-%5Cfrac%7Bx%5E3%7D%7B2%7D%7D%20%3D-%5Cfrac%7B2%7D%7B3%7D%20

题(5):求极限%5Clim_%7Bx%20%5Cto%200%7D%20%5B%5Cfrac%7B1%7D%7B%5Cln%20(x%2B%5Csqrt%7B1%2Bx%5E2%7D%20)%7D-%5Cfrac%7B1%7D%7B%5Cln%20(1%2Bx)%2B%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20%7D%20%20%5D

莫慌,先将函数通分:

%5Clim_%7Bx%20%5Cto%200%7D%20%5B%5Cfrac%7B%5Cln%20(1%2Bx)%2B%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20-%5Cln%20(x%2B%5Csqrt%7B1%2Bx%5E2%7D%20)%7D%0A%7B%5Cln%20(x%2B%5Csqrt%7B1%2Bx%5E2%7D%20)(%5Cln%20(1%2Bx)%2B%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20)%7D%20%20%5D

其中ln(1+x)和ln[x+√(x²+1)]的泰勒展开我们在前两篇文章已经推导过了,附上当复习:

%5Cbbox%5B%238fffa8%5D%7B%5Csmall%20%5Cbegin%7Balign%7D%0A%5Cln%20(1%2Bx)%26%3Dx-%5Cfrac%7Bx%5E2%7D%7B2%7D%2B%5Cfrac%7Bx%5E3%7D%7B3%7D-%5Ccdots%20%2B%5Cfrac%7B(-1)%5E%7Bn-1%7D%7D%7Bn%7Dx%5En%2B%5Comicron%20(x%5En)%5C%5C%0A%5Cln%20(x%2B%5Csqrt%7B1%2Bx%5E2%7D%20)%26%3Dx-%5Cfrac%7B1!!%7D%7B2!!%5Ccdot%203%7D%20%20x%5E3%2B%5Cfrac%7B3!!%7D%7B4!!%5Ccdot%205%7Dx%5E5-%5Ccdot%20%2B%0A%5Cfrac%7B(-1)%5En%20(2n-1)!!%7D%7B(2n)!!(2n%2B1)%7Dx%5E%7B2n%2B1%7D%2B%5Comicron%20(x%5E%7B2n%2B1%7D)%20%20%20%20%0A%5Cend%7Balign%7D%7D%20

先对分母两个因式进行估阶:

其中ln[x+√(1+x²)]=x+o(x)~x

%5Cln%20(1%2Bx)%2B%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20进行估阶

其中ln(1+x)=x+o(x)~x,为1阶;

%5Cbegin%7Balign%7D%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%26%20%3D%5Cint_%7B0%7D%5E%7Bx%7Dt%5B%5Cmathrm%7Be%7D%20%2B%5Comicron%20(1)%5D%5Cmathrm%7Bd%7Dt%5C%5C%0A%26%3D%5Cint_%7B0%7D%5E%7Bx%7D%5B%5Cmathrm%7Be%7Dt%20%2B%5Comicron%20(t)%5D%5Cmathrm%7Bd%7Dt%5C%5C%0A%26%3D%5Cfrac%7B%5Cmathrm%7Be%7D%20%7D%7B2%7D%20x%5E2%2B%5Comicron%20(x%5E2)%0A%5Cend%7Balign%7D

阶数为2


进而

%5Cbegin%7Balign%7D%0A%26%5Cln%20(1%2Bx)%2B%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20%5C%5C%0A%3D%26x%2B%5Comicron%20(x)%2B%5Cfrac%7B%5Cmathrm%7Be%7D%20%7D%7B2%7D%20x%5E2%2B%5Comicron%20(x%5E2)%5C%5C%0A%3D%26x%2B%5Comicron%20(x)%5C%5C%0A%5Csim%20%26x%0A%5Cend%7Balign%7D

则分母~x*x=x²,阶数为2,故分子也需展开到2阶

准则(1),则分子的3项各自都需要展开到2阶

其中ln(1+x)=x-x²/2+o(x²)

ln[x+√(1+x²)]=x+o(x²)

%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20是变上限积分,由准则(4)知,需要对被积函数展开到2-1=1阶

而被积函数t*(1+t)^(1/t)又是函数乘积,t为一阶,则(1+t)^(1/t)需展开到1-1=0阶

你瞧,根据这4个选阶准则,咱们做极限题就像剥洋葱一样把每一部分需要展开的阶数逐步确定下来了,以后就再也不用愁漏阶啦~

于是逐步回推得:

%5Cbegin%7Balign%7D%0A%26(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%3D%5Cmathrm%7Be%7D%20%2B%5Comicron%20(1)%5C%5C%0A%5CRightarrow%20%26t(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%3D%5Cmathrm%7Be%7D%20t%2B%5Comicron%20(t)%5C%5C%0A%5CRightarrow%20%26%5Cint_%7B0%7D%5E%7Bx%7D%20t(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%5Cmathrm%7Bd%7Dt%20%3D%5Cfrac%7B%5Cmathrm%7Be%7D%20%7D%7B2%7D%20x%5E2%2B%5Comicron%20(x%5E2)%0A%5Cend%7Balign%7D

进而有:

%5Cbegin%7Balign%7D%0A%26%5Cln%20(1%2Bx)%2B%0A%5Cint_%7B0%7D%5E%7Bx%7Dt(1%2Bt)%5E%7B%5Cfrac%7B1%7D%7Bt%7D%20%7D%20%5Cmathrm%7Bd%7Dt%20-%5Cln%20(x%2B%5Csqrt%7B1%2Bx%5E2%7D%20)%5C%5C%0A%3D%26%5Bx-%5Cfrac%7Bx%5E2%7D%7B2%7D%2B%5Comicron%20(x%5E2)%5D%2B%5B%5Cfrac%7B%5Cmathrm%7Be%7D%20%7D%7B2%7D%20x%5E2%2B%5Comicron%20(x%5E2)%5D-%5Bx%2B%5Comicron%20(x%5E2)%5D%20%5C%5C%0A%3D%26%5Cfrac%7B%5Cmathrm%7Be%7D-1%20%7D%7B2%7D%20x%5E2%2B%5Comicron%20(x%5E2)%20%5C%5C%0A%5Csim%20%26%5Cfrac%7B%5Cmathrm%7Be%7D-1%20%7D%7B2%7D%20x%5E2%0A%5Cend%7Balign%7D

最终得:

%5Ctext%7B%E5%8E%9F%E5%BC%8F%7D%3D%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cfrac%7B%5Cmathrm%7Be%7D%20-1%7D%7B2%7Dx%5E2%20%7D%7Bx%5E2%7D%20%3D%5Cfrac%7B%5Cmathrm%7Be%7D%20-1%7D%7B2%7D

题(6):求极限%5Clim_%7Bx%20%5Cto%200%7D%20%5Cint_%7B0%7D%5E%7Bx%7D%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%5Ccot%20x%5Cmathrm%7Bd%7Dt%20

这变限积分嵌套的形式,狗看了都摇头啊...

莫慌,咋们一步步来拆解~

首先最外面是对t积分,那么变限积分就是一个关于t的函数,因此cotx对关于t变限积分而言就是与t无关的常数,直接提到积分号外面。并且取倒数得1/tanx,再利用tanx~x初步化简为:

%5Cbegin%7Balign%7D%0A%5Ctext%7B%E5%8E%9F%E5%BC%8F%7D%26%3D%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cint_%7B0%7D%5E%7Bx%7D%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%5Cmathrm%7Bd%7Dt%20%7D%7B%5Ctan%20x%7D%5C%5C%0A%26%3D%20%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cint_%7B0%7D%5E%7Bx%7D%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%5Cmathrm%7Bd%7Dt%20%7D%7Bx%7D%0A%5Cend%7Balign%7D

分母是1阶,那么分子也需要展到1阶。由准则(4),被积函数需要展开到1-1=0阶。这说明我们只需要求出极限%5Clim_%7Bt%20%5Cto%200%7D%20%5C%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D即可

接下来求这个极限。观察这个极限,发现是1^∞型未定式,由前文提及的“三部曲”即得:

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bt%20%5Cto%200%7D%20%5C%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bt%20%5Cto%200%7D%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D-1)%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bt%20%5Cto%200%7D(%5Cfrac%7B%5Carctan%20t-t%7D%7Bt%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%7D)%5D%0A%5Cend%7Balign%7D

分子用arctant的泰勒展开估阶,分母用前文提及的变限积分等价的结论,即得:

%5Cbegin%7Balign%7D%0A%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bt%20%5Cto%200%7D%5Cfrac%7B%5Carctan%20t-t%7D%7Bt%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%7D%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bt%20%5Cto%200%7D%5Cfrac%7Bt-%5Cfrac%7Bt%5E3%7D%7B3%7D%2B%5Comicron%20(t%5E3)%20-t%7D%7Bt%5Cint_%7B0%7D%5E%7Bt%7D(u%2B%5Comicron%20(u))%5Cmathrm%7Bd%7Du%7D%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bt%20%5Cto%200%7D%5Cfrac%7B-%5Cfrac%7Bt%5E3%7D%7B3%7D%2B%5Comicron%20(t%5E3)%7D%7Bt%5Ccdot(%20%5Cfrac%7B1%7D%7B2%7Dt%5E2%20%2B%5Comicron%20(t%5E2))%7D%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20%5B%5Clim_%7Bt%20%5Cto%200%7D%5Cfrac%7B-%5Cfrac%7Bt%5E3%7D%7B3%7D%2B%5Comicron%20(t%5E3)%7D%7B%5Cfrac%7B1%7D%7B2%7Dt%5E3%2B%5Comicron%20(t%5E3)%20%7D%5D%5C%5C%0A%3D%26%5Cmathrm%7Bexp%7D%20(-%5Cfrac%7B2%7D%7B3%7D%20)%0A%5Cend%7Balign%7D


逐步回推即得:

%5Cbegin%7Balign%7D%0A%26%5Clim_%7Bt%20%5Cto%200%7D%20%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%3D%5Cmathrm%7Be%7D%20%5E%7B-%5Cfrac%7B2%7D%7B3%7D%20%7D%5C%5C%0A%5CRightarrow%20%26(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%3D%5Cmathrm%7Be%7D%20%5E%7B-%5Cfrac%7B2%7D%7B3%7D%20%7D%2B%5Comicron%20(1)%5C%5C%0A%5CRightarrow%20%26%5Cint_%7B0%7D%5E%7Bx%7D%20(%5Cfrac%7B%5Carctan%20t%7D%7Bt%7D%20)%5E%7B%5Cfrac%7B1%7D%7B%5Cint_%7B0%7D%5E%7Bt%7D%5Cln%20(1%2Bu)%5Cmathrm%7Bd%7Du%20%7D%20%7D%5Cmathrm%7Bd%7Dt%20%0A%3D%5Cmathrm%7Be%7D%20%5E%7B-%5Cfrac%7B2%7D%7B3%7D%20%7Dx%2B%5Comicron%20(x)%5C%5C%0A%5CRightarrow%20%26%5Ctext%7B%E5%8E%9F%E5%BC%8F%7D%3D%5Clim_%7Bx%20%5Cto%200%7D%20%5Cfrac%7B%5Cmathrm%7Be%7D%20%5E%7B%5Cfrac%7B2%7D%7B3%7D%20%7Dx%2B%5Comicron%20(x)%7D%7Bx%7D%20%0A%3D%5Cmathrm%7Be%7D%20%5E%7B-%5Cfrac%7B2%7D%7B3%7D%20%7D%0A%5Cend%7Balign%7D


利用估阶的核心思想,我们见招拆招很快就解决了这样一个庞然大物!而估阶的核心思想是什么?就是笔者帮大家总结了的那4种选阶准则(分别用于处理函数加减和数乘函数乘法函数复合函数求导/求积的展开)

总结

本篇文章补充了渐进曲线的概念,并讲解了渐进曲线的求法(其实就是复合函数泰勒展开的一种特殊情况罢了),由此一部分x→∞时的极限依然可用泰勒展开求解。

然后带大家练习了几道极限题目,其中还顺带分享了几个小结论:

(1)1^∞型幂指函数极限讲了一个“三部曲”(先化复合函数,再交换复合与求极限次序,最后等价无穷小),利用“三部曲”即可转化为求一个∞*0型未定式极限

(2)变上限积分等价无穷小:被积函数和积分上限分别等价

(3)极限题一般优先考虑对函数进行因式分解。例子如下:

遇到sin()-sin(),cos()-cos()的因式,考虑和差化积会相对方便;

遇到ln()-ln()的因式,考虑对数运算进行合并;

遇到exp()-exp()的因式,考虑利用eᵃ-eᵇ=eᵇ(eᵃ⁻ᵇ-1)化为一个非零因子和一个零因子,非零因子直接代入,零因子用等价化为a-b,进而简化因式


水了超8000字聊~从题库找到不少的极限题,90%以上都可以用估阶的思想解决,如果后续还对其保持上头,那么可能会再水一篇文章以估阶思想为武器打极限小怪~

看来开学近半个月了大家都很忙,也没图啥流量就当是自己整理笔记好了~



Copyright © 2024 aigcdaily.cn  北京智识时代科技有限公司  版权所有  京ICP备2023006237号-1