INEQUALITY PROOF
来源:互联网 发布:动漫下载软件 编辑:程序博客网 时间:2024/05/20 04:14
INEQUALITY PROOF
Ethan
In Loving Memory of Mamba Day, 4.13, 2016
I. Common Inequalities
(a). Lagrangian Formula for Summation Estimation
Suppose f(x) is a continuous function on x >= 1 and F(x) is its primitive function. Iff(x) is monotonically decreasing,
The inequality is reversed iff(x) is monotonically increasing.
Proof:
Suppose f(x) is monotonically increasing.F’(x) = f(x). By applying Lagrange mean value theorem toF(x) on [k, k+1], we have
Note that
By summing up these inequalities above, we obtain,
Therefore,
Similarly, we can prove the other case in whichf(x) is monotonically decreasing.
(b). The r-order Inequality
The r-order meanMr(a, q) is defined as
where a = (a1, …, an)T > 0 , q = (q1, …, qn)T andq1+…+qn = 1.
Theorem:Mr(a, q) is a monotonically increasing function on R with respect tor.
Proof:
Let bi =air,i= 1, …,n,
where B is a random variable with a generalized Bernoulli distribution,
Note that f(x) = xlnx is a convex function. Thus, the expectation of a function value is larger than the function value of the expectation (See Appendix I), i.e.,
This completes the proof.#
Corollary I:
When qi=1/n, we have,
Arithmetic mean (r= 1),
Geometric mean (r = 0),
Harmonic mean (r = -1),
H(a)<=G(a)<=A(a).
Corollary II (Holder Inequality)
Suppose we havem vectors with n components,
andw1+w2+…+wm= 1, wi >0, then
A relaxation on its left hand side and some changes of variables might give us more insight on the Holder inequality, i.e.,
letting rijwi =bij, andpi = 1/wi,
By defining a generalized “inner product” among several vectors,
we obtain a relaxed Holder inequality,
wherepi’s are called conjugate coefficients satisfying
Proof:
This completes the proof.#
II. Integral Inequalities
(a). The r-order inequality
In this section, we are going to consider various means of a positive-valued continuous functionf(x) on [a, b]. The interval [a, b] can be evenly divided inton sub-intervals with a=x0<x1<…<xn=b, wherexk=a+k(b-a)/n, k=0, 1, …,n.
Arithmetic mean of f(x) on [a, b],
Geometric mean of f(x) on [a,b],
Harmonic mean of f(x) on [a,b],
In generalization, ther-order mean off(x) on [a,b] is given by
Note that A(f)=M1(f),G(f)=M0(f) and H(f)=M-1(f). One can also show that Mr(f) is a monotonically increasing function wrt r. In particular,H(f) <= G(f) <= A(f).
(b) Common Methods to Prove Integral Inequalities
(b1) Holder Inequality
Given m positive-valued continuous functions on [a,b], f1(x), f2(x), …,fm(x), and positive scalars w1, …,wm satisfying w1+…+wm=1, then
Similarly, we have a relaxed Holder inequality
where the generalized inner product among functions is given by
andpi’s are called conjugate coefficients satisfying
Particularly, for a “two functions andw1=w2=1/2” case,
or equivalently,
We obtain the Cauchy-Schwarz Inequality (See Appendix II).
(b2) Function Convexity
Definition (Convex function):f(x) is a convex function if and only if
for
Particularly, whenalphai=1/n, we obtain the Jensen Inequality.
One can show that
If f(x) is differentiable,f(x) is convex <=> f’(x) monotonically increases;
If f(x) is twice differentiable,f(x) is convex <=> f’’(x)< 0.
Corollary: Supposef(x) is an integrable function on [a, b], m<=f(x)<=M, g(x) is a continuous convex function on [m, M], then
Hint: This can be proved by taking a limit on Jensen Inequality.
APPENDIX
I. A Property of Convex Function
Prove: for a convex twice differentiable real-valued functionf(x) defined on [a, b], the expectation of f(x) is equal to or larger than the function value of the expectation.
Actually this property directly follows the definition of a convex function by regarding alpha’s as probabilities or probability density. Here is another proof when the convex function is twice differentiable.
Proof:
Since the second derivative off is nonnegative, its first derivative must be nondecreasing. Using the fundamental theorem of calculus, we obtain
Now suppose a random variableX can take every possible valuex. Choosing c =E[X] and taking expectations on both sides, we have
This completes the proof.#
II. Cauchy-Schwarz Inequality
(1) Generic Cauchy-Schwarz Inequality
Recall that, in functional analysis,
In an inner-product space (H, K,<.,.>), for anyx and y in X, the generic Cauchy-Schwarz inequality holds as
with the equality iff x and y are linearly dependent.
(2) Variants of Cauchy-Schwarz Inequality
There are myriads variants of the Cauchy-Schwarz inequality.
(a) The inner-product space ofm-by-n matrices
Suppose H = {m-by-n complex matrices}, K = C, and the inner product is defined as <A, B> = trace(A*B) (One can check the validity of this inner product in (H, K, <.,.>) by examining the definition of the general inner product). Then, the Cauchy-Schwarz Inequality holds as
(b) The inner-product space of random variables
Suppose H = {random variables}, K = R, and the inner product is defined as <X,Y> = Cov(X, Y) (One can check the validity of this inner product in (H, K, <.,.>) by examining the definition of the general inner product). Then, the Cauchy-Schwarz inequality holds as
(c) The inner-product space of random matrices
Suppose A andB are p-by-n and q-by-n random matrices such that E||A||2< infinity, E||B||2 < infinity and E(AAT) is nonsingular. Then, (1)
In particular, whenn=1,A:=X-EX and B:=Y-EY, we have (2)
Proof:
Let LAMBDA = E(BAT)E-1(AAT). Stylistically,LAMBDA is a linear operator sequentially performing a whitening transform (eliminating the linear independence ofA’s components) and a projection onto the space of B. Then,
This completes the proof.#
Pictorially, Figure 1 shows the transformation of (2).
Philosophically, Formula (1) and (2) are both high-dimensional Cauchy-Schwarz Inequality, regarding a transformation between linear spaces and one between vectors, respectively.
- INEQUALITY PROOF
- Triangle inequality
- String inequality
- H?lder\'s inequality
- Erdos–Mordell inequality
- VC Theory: Hoeffding Inequality
- Inequality of chinese education
- Hoeffding inequality and bound
- Yeahpeng's inequality
- Inequalities - Holder's inequality
- Inequalities - Minkowski's inequality
- Inequalities - Tchebychef's inequality
- An inequality about sine function
- 马尔可夫不等式 Markov's inequality
- Chebyshev-Inequality-with-Linear-Bound
- EAS & Theft-Proof Series
- bullet proof vests
- 防呆法(Fool Proof)
- 行业很重要
- Activiti源码分析
- OC与swift混编
- 其他形式的链表
- javaweb学习总结(四十五)——监听器(Listener)学习二
- INEQUALITY PROOF
- 小白排序之插入排序
- 纯干货!面试官如何面试程序员
- 1. Two Sum
- VMware下ubuntu扩展磁盘空间
- 使用自定义模板为Myeclipse添加新建Filter的功能
- 将本地代码push到GitLab空仓库
- 抽象类和接口联系与区别
- Designing for Multi-Window