若f(x)在x0的某一领域U内连续,且下列极限存在,则f(x)在x0处可导.
- x→x0limx−x0f(x)−f(x0)
- Δx→0limΔxf(x0+Δx)−f(x0)
eg: 设f(x)在区间[−δ,δ]内有定义,f(0)=1,且满足:
x→0limx2ln(1−2x)+2xf(x)=0
证明f(x)在x=0处可导,并求f′(0).
解:
原式=x→0limx2−2x−21(−2x)2+2xf(x)=x→0limx22xf(x)−2x−2=2x→0limxf(x)−1−2=0
∴x→0limxf(x)−1=1=x→0limx−0f(x)−f(0)=f′(0)
∴f′(0)=1
下式中的a均为常数
(xa)′=axa−1(a)
(ax)′=axlna(a>0,a̸=1)
(ex)′=(ex)
(logax)′=xlna1(a>0,a̸=1)
(lnx)′=x1
(sinx)′=cosx
(cosx)′=−sinx
(tanx)′=(cosxsinx)′=cos2xcos2x+sin2x=cos2x1=sec2x
(cotx)′=(sinxcosx)′=sin2x−sin2x−cos2x=−sin2x1=−csc2x
(secx)′=(cosx1)′=cos2x0+sinx=tanxsecx
(cscx)′=(sinx1)′=sin2x0−cosx=−cotxcscx
(arcsinx)′=1−x21
(arccosx)′=−1−x21
(arctanx)′=1+x21
(arccotx)′=−1+x21
(ln(x+a2±x2))′=a2±x21 🌠
(sinax)(n)=ansin(ax+2π.n)
(cosax)(n)=ancos(ax+2π.n)
(ax)(n)=ax(lna)n
(eax)(n)=aneax
(lnx)(n)=(−1)n−1⋅xn(n−1)!,(x>0)
[ln(x+1)](n)=(−1)n−1⋅(x+1)n(n−1)!,(x>−1)
(x+a1)(n)=(−1)n⋅(x+a)n+1n!
(u⋅v)(n)=k=0∑nCnku(n−k)v(k)=u(n)v+nu(n−1)v′+2n(n−1)u(n−2)v′′+⋯+uv(n)
设f(x)为连续函数,φ1(x),φ2(x)均可导,则有:
(⎰φ1(x)φ2(x)f(t)dt)′=f(φ2(x))φ2′(x)−f(φ1(x))φ1′(x)
设函数y=f(x)由某一参数式f(x)={x=x(t)y=y(t) 确定,并设x(t),y(t)均可导,x′(t)̸=0, 则:
yx′=xt′yt′
yxx′′=xt′(yx′)t′
设函数y=f(x)由方程F(x,y)=0确定,视F(x,y)中的y为x的函数f(x), 对方程F(x,y)=0两边求导,得到一个关于yx′的式子, 解出yx′即可, 在分母不为0的情况下,将上面求过一次导数的式子进一步对x求导, 得到一个同时包含yx′,yx′′的式子, 代入上一步中求得的yx′,即可得到yx′′.
u(x)v(x)=ev(x)lnu(x)
(u(x)v(x))′=(ev(x)lnu(x))′=ev(x)lnu(x)[v′(x)lnu(x)+u(x)v(x)u′(x)]=u(x)v(x)[v′(x)lnu(x)+u(x)v(x)u′(x)]
设 y=f(x)可导且f′(x)̸=0, 则存在反函数x=φ(y),且dydx=dxdy1=f′(x)1, 即φ′(y)=f′(x)1
若y=f(x)存在二阶导数, 则φ′′(y)=dydφ′(y)=dyd(f′(x)1)=dxd(f′(x)1)⋅dydx=(f′(x))20−f′′(x)⋅f′(x)1=−(f′(x))3f′′(x)
反函数存在的充分条件: 设y=f(x)在区间(a,b)内为严格单调的连续函数,则它必存在具有相同单调性的严格单调反函数.