关于如何使用好复合函数求导公式的探讨

复合函数的求导问题,历来是函数求导数中的一个难点.关于复合函数的求导法则,国内、国外的数学分析教材和高等数学版本都是这样叙述的:设y=f[（?）(x)]是由函数y=f(u)及u=（?）(x)复合而成的函数,若函数u=（?）(x)在点x处是可导的,y=f(u)在对应点y=（?）(x)处也可导,则复合函数y=f[（?）(x)]在点x处可导,且其导数为     

多元复合函数求导链式法则证明的注记

"当自变量增量趋向于0时,中间变量增量可能取到0"是多元复合函数求导链式法则证明过程中的主要困难.本文通过补充定义的方法克服该难点,在外层函数可微、内层函数可导这一最弱条件下严格化了证明.     

求多元函数二阶偏导数的矩阵方法

多元函数求偏导问题是多元函数微分学中的一项重点和难点内容。在求解这类题目时 ,既要严格区分自变量与中间变量 ,而且要注意不能丢掉偏导函数作为复合函数时的偏导数。特别求二阶偏导时 ,学生容易漏项 ,有没有比较好的方法 ?先考察下例 :例 1　 u =f ( x +y,xy,xyz) ,求 2 ux2解　设 t=x +y,v =xy,w =xyz,则 u =f ( t,v,w) ,按照多元复合函数求导法则求导如下 :ux=ft+fv. y +fw. yz =f′1+yf′2 +yzf′3　　　 2 ux2 =f″11+f″12 . y +f″13 . yz +yf″2 1+yf″2 2 . y +yf″2 3 . yz +yzf″3 1+yzf″3 2 . y +y…     

借助函数图形加深对定理的理解

利用函数图形解释复合函数的极限运算法则及其证明思路、由参数方程所确定的函数的导数以及复合函数求导法则、柯西中值定理的证明思路。     

复合函数求导法则的又一证明

复合函数的求导法则是求导运算的重要法则;对于ｙ＝ｆ（ｕ）,ｕ＝ｇ（ｘ）,复合函数ｙ＝ｆ〔ｇ（ｘ）〕的求导法则的证明有一个很自然的想法：ΔｙΔｘ＝ΔｙΔｕ·ΔｕΔｘ,ｌｉｍΔｘ→０ΔｙΔｘ＝ｌｉｍΔｕ→０ΔｙΔｕ·ｌｉｍΔｘ→０ΔｕΔｘ;但是,当Δｘ→０时,Δｕ可能等于０,此时ΔｙΔｕ没有意义,所以上面很直接的想法行不通;一般的证明采取另外的方法［１］,［２］;本文仍从上面直观的想法出发,加以改进,得到了又一个证明;定理　若ｙ＝ｆ（ｕ）在ｕ可导,函数ｕ＝ｇ（ｘ）在ｘ可导,则复合函数ｙ＝ｆ〔ｇ（ｘ）…     

复合函数求导链式法则的一个注记

本文考虑在复合函数求导链式法则的证明过程中,增加"当Δu=0时,定义α=0"的必要性,并给出一个反例.     

关于分段函数及含绝对值函数的求导问题

本文讨论分段函数的求导问题，建立了求导时方法选取的一般程式。对于含绝对值的函数，给出了一个求导定理。一、分段函数的导数分段函数的求导，关键在于求分段点处的导数，常用方法有：①不连续则不可导；②导数或左右导数的定义；③导数单侧极限定理*：设f（x）在（a，b）内连续，x0∈（a，b），在（a，x0）及（x0，b）内可导且limf（x）、limf（x）都存在，则导数单侧极限定理用左右导数定义及微分中值定理可证，此处从略。下面仅作几点说明：1“定理中若厂十（X。）一片一（X。），则几X）在X。处可导，若不相等，则人X）在X。处不…     

多元多重复合函数的求导法则

<正> 大家熟知,一元复合函数y=f(u)u=φ(v)v=ψ(x)的求导公式是     

幂指函数的求导法则

<正> 形如y=u(x)~(v(x))(u(x)>0)的函数称为幂指函数。幂指函数属于初等函数的一部分,它是微分学中经常迂到的函数类型之一。求幂指函数的导数通常有两种方法,一是利用“对数求导法”,也即先对函数表达式取对数,然后再求导数;二是利用对数恒等式将幂指函数化为以e为底的复合指数函数形式,再求其导数,本质上都是用复合函数求导法则。先看下面的例子:     

序轴法--复合函数单调区间的一种简捷求法

复合函数是高中数学中的一类重要函数 ,讨论复合函数的单调性 ,求出其单调区间是复合函数问题中的一类重要问题 .本文介绍一种求复合函数单调区间的简捷方法 ,供大家参考 .本文介绍的复合函数单调区间求法的理论依据是下面的定理 (判定定理 )　若 y =F1(x) ,u1=F2 (x) ,… ,un=Fn 1(x)都是单调函数 ,则 n次复合函数 y =F1{ F2 [… Fn 1(x) ]}在其定义域内也是单调函数 ,且它为增函数的充要条件是 y =F1(x) ,u1=F2 (x) ,… ,un =Fn 1(x)中减函数的个数为偶数 ;它为减函数的充要条件是y =F1(x) ,u1=F2 (x) ,… ,un=Fn 1(x)中减函数的个数…     

Chain rule for conic derivatives

For all “nice” definitions of differentiability, the Chain Rule should be valid. We show that the Chain Rule remains true for some wide class of definitions of differentiability if one considers as approximative mappings (derivatives) not just continuous linear, but positively homogeneous mappings satisfying certain topological conditions (which are fulfilled for continuous linear mappings). For brevity, we call such derivatives conic. We will give corollaries for the case of conic differentiation of mappings between normed spaces, especially for the case of Fréchet conic differentiation and compact conic differentiation.     

K-quasiderivations

A K-quasiderivation is a map which satisfies both the Product Rule and the Chain Rule. In this paper, we discuss several interesting families of K-quasiderivations. We first classify all K-quasiderivations on the ring of polynomials in one variable over an arbitrary commutative ring R with unity, thereby extending a previous result. In particular, we show that any such K-quasiderivation must be linear over R. We then discuss two previously undiscovered collections of (mostly) nonlinear K-quasiderivations on the set of functions defined on some subset of a field. Over the reals, our constructions yield a one-parameter family of K-quasiderivations which includes the ordinary derivative as a special case.     

具有变时滞的离散双向联想记忆神经网络的全局指数稳定性

利用Lyapunov函数和线性矩阵不等式(LMI),给出了判定一类双向联想记忆(BAM)神经网络模型的指数稳定的充分性条件.该条件去掉了以往论文中所要求的激活函数单调,可微分的条件,而且所得结果利用里的工具易于检测.并举例说明本文结果的有效性.     

Lower tail dependence for Archimedean copulas: Characterizations and pitfalls

Tail dependence copulas provide a natural perspective from which one can study the dependence in the tail of a multivariate distribution. For Archimedean copulas with continuously differentiable generators, regular variation of the generator near the origin is known to be closely connected to convergence of the lower tail dependence copulas to the Clayton copula. In this paper, these characterizations are refined and extended to the case of generators which are not necessarily continuously differentiable. Moreover, a counterexample is constructed showing that even if the generator of a strict Archimedean copula is continuously differentiable and slowly varying at the origin, then the lower tail dependence copulas still do not need to converge to the independent copula.     

Jacobi矩阵链法则中的补充证明

本文指出 BOOTHBY W M的《微分流形与黎曼几何引论》(见文献 [1 ])一书中关于 Jacobi矩阵的链法则的处理中存在的问题 ,并对其补充证明     

A class of practically nowhere differentiable functions on the complex plane

This note could find classroom use in an introductory course on complex analysis. Using some of the most significant theorems from complex analysis, the main result provides a simple method for transforming many elementary functions (defined on the complex plane) into everywhere continuous functions that are differentiable only on a nowhere dense set. Accordingly, such continuous functions are termed ‘practically nowhere differentiable’. The twofold pedagogical value of this method is that (1) students can readily generate examples of everywhere continuous, practically nowhere differentiable functions that do not require any direct appeal to infinite series, and (2) the often dynamical difference between the behaviour of functions of a complex variable and functions of a real variable is showcased.     

On strong pseudomonotonicity and (semi)strict quasimonotonicity

New concepts of strong pseudomonotonicity, strict quasimonotonicity, and semistrict quasimonotonicity of a map are introduced and their properties are studied. In the case of a differentiable gradient map, we show that strong pseudomonotonicity of the gradient is equivalent to strong pseudoconvexity of the underlying function. This does not hold for a different concept of strong pseudomonotonicity in Ref. 1. Analogous results are shown for strict quasimonotonicity and semistrict quasimonotonicity.     

Sensitivity analysis framework for variational inequalities

In this paper a sensitivity analysis framework is developed for variational inequalities. The perturbed solution to a parametric variational inequality problem is shown to be continuous and directionally differentiable under appropriate second order and regularity assumptions. Moreover, this solution is once continuously differentiable if strict complementarity also holds.     

Existence et determination du premier point de bifurcation pour des couples d'opérateurs potentiels non différentiables à l'origine

Summary Let Λ and Γ denote respectively the gradients of two Gateaux differentiable real functionals ϕ and ϕ on a real reflexive Banach space V. We shall prove that, under very general assumptions, the first bifurcation point (at the origin in V) for the non-linear equation Γ(u)=λΛ(u) is the first eigenvalue of the linear equation B(u)=λA(u) where Λ=A+ν and Γ=B+ω. The perators Λ and Γ are not necessarily differentiable at the origin. Roughly speaking it sufficies to assume that the projections in a suitable direction of the remainders ν(u) and ω(u) go to zero faster than ∥u∥; this direction depends on u ans is, in the simplest examples, actually the u-direction.

---

---

Entrata in Redazione il 26 ottobre 1977.     

Plane-Like Minimizers and Differentiability of the Stable Norm

In this paper we investigate the strict convexity and the differentiability properties of the stable norm, which corresponds to the homogenized surface tension for a periodic perimeter homogenization problem (in a regular and uniformly elliptic case). We prove that it is always differentiable in totally irrational directions, while in other directions, it is differentiable if and only if the corresponding plane-like minimizers satisfying a strong Birkhoff property foliate the torus. We also discuss the issue of the uniqueness of the correctors for the corresponding homogenization problem.