### 统计代写|最优控制作业代写optimal control代考|Singularities of Semiconcave Functions

statistics-lab™ 为您的留学生涯保驾护航 在代写最优控制optimal control方面已经树立了自己的口碑, 保证靠谱, 高质且原创的统计Statistics代写服务。我们的专家在代写最优控制optimal control代写方面经验极为丰富，各种代写最优控制Soptimal control相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• (Generalized) Linear Models 广义线性模型
• Statistical Machine Learning 统计机器学习
• Longitudinal Data Analysis 纵向数据分析
• Foundations of Data Science 数据科学基础

## 统计代写|最优控制作业代写optimal control代考|Singularities of Semiconcave Functions

By a singular point, or singularity, of a semiconcave function $u$ we mean a point where $u$ is not differentiable. This chapter is devoted to the analysis of the set of all singular points for $u$, which is called singular set and is denoted here by $\Sigma(u)$. As we have already remarked, the singular set of a semiconcave function has zero measure by Rademacher’s theorem. However, we will see that much more detailed properties can be proved.

In Section $4.1$ we study the rectifiability properties of the singular set. We divide the singular points according to the dimension of the superdifferential of $u$ denoting by $\Sigma^{k}(u)$ the set of points $x$ such that $D^{+} u(x)$ has dimension $k$. Then we show that $\Sigma^{k}(u)$ is countably $(n-k)$-rectifiable for all integers $k=1, \ldots, n$. In particular, the whole singular set $\Sigma(u)$ is countably ( $n-1)$-rectifiable.

Sections $4.2$ and $4.3$ are devoted to the propagation of singularities for semiconcave functions: given a singular point $x_{0}$, we look for conditions ensuring that $x_{0}$ belongs to a connected component of dimension $v \geq 1$ of the singular set. We study first the propagation along Lipschitz arcs and then along Lipschitz manifolds of higher dimension. In general we find that a sufficient condition for the propagation of singularities from $x_{0}$ is that the inclusion $D^{*} u\left(x_{0}\right) \subset \partial D^{+} u\left(x_{0}\right)$ (see Proposition 3.3.4-(b) ) is strict.

As an application of the previous analysis, we study in Section $4.4$ some properties of the distance function from a closed set $S$. Using our propagation results, we show that the distance function has no isolated singularities except for the special case when the singularity is the center of a spherical connected component of the complement of $S$. In general, we show that a point $x_{0}$ which is singular for the distance function belongs to a connected set of singular points whose Hausdorff dimension is at least $n-k$, with $k=\operatorname{dim}\left(D^{+} d\left(x_{0}\right)\right)$.

## 统计代写|最优控制作业代写optimal control代考|Rectifiability of the singular sets

Throughout this chapter $\Omega \subset \mathbb{R}^{n}$ is an open set and $u: \Omega \rightarrow \mathbb{R}$ is a semiconcave function. We denote by $\Sigma(u)$ the set of points of $\Omega$ where $u$ is not differentiable and

we call it the singular set of $u$. In the following we use some notions from measure theory, like Hausdorff measures and rectifiable sets, which are recalled in Appendix A. 3 .

We know from Theorem 2.3.1-(ii) that $D u$ is a function with bounded variation if $u$ is semiconcave with a linear modulus. For functions of bounded variation one can introduce the jump set, whose rectifiability properties have been widely studied (see Appendix A. 6). We now show that the jump set of $D u$ coincides with the singular set $\Sigma(u)$. To this purpose we need two preliminary results. The first one is a lemma about approximate limits (see Definition A. 6.2).

Lemma 4.1.1 Let $w \in L^{1}(A)$, with $A \subset \mathbb{R}^{n}$ open, let $\bar{x} \in A$, and let ap $\lim {x \rightarrow \bar{x}} w(x)=\bar{w}$. Then for any $\theta \in \mathbb{R}^{n}$ with $|\theta|=1$ we can find a sequence $\left{x{k}\right} \subset A$ such that
$$x_{k} \rightarrow \bar{x}, \quad \frac{x_{k}-\bar{x}}{\left|x_{k}-\bar{x}\right|} \rightarrow \theta, \quad w\left(x_{k}\right) \rightarrow \bar{w} \quad \text { as } k \rightarrow \infty$$
Proof – For any $k \in \mathbb{N}$, let us define
$$A_{k}=\left{x \in A \backslash{\bar{x}}:\left|\frac{x-\bar{x}}{|x-\bar{x}|}-\theta\right|<\frac{1}{k}\right}$$ Any such set $A_{k}$ is the intersection of $A$ with an open cone of vertex $\bar{x}$. Therefore $$\lim {\rho \rightarrow 0^{+}} \frac{\text { meas }\left(B{\rho}(\bar{x}) \cap A_{k}\right)}{\rho^{n}}>0 .$$
By the definition of approximate limit we have
$$\lim {\rho \rightarrow 0^{+}} \frac{\operatorname{meas}\left(\left{x \in B{\rho}(\bar{x}) \cap A_{k}:|w(x)-\bar{w}|>1 / k\right}\right)}{\rho^{n}}=0 .$$
Comparing the above relations we see that the set
$$\left{x \in B_{\rho}(\bar{x}) \cap A_{k}:|w(x)-\bar{w}| \leq 1 / k\right}$$
is nonempty if $\rho$ is small enough. Thus, we can find $x_{k} \in A_{k}$ such that $\left|w\left(x_{k}\right)-\bar{w}\right| \leq$ $1 / k,\left|x_{k}-\bar{x}\right| \leq 1 / k$. Repeating this construction for all $k$ we obtain a sequence $\left{x_{k}\right}$ with the desired properties.

Next we give a result showing, roughly speaking, that for the gradient of a semiconcave function the notions of limit and of approximate limit coincide.

## 统计代写|最优控制作业代写optimal control代考|Propagation along Lipschitz arcs

Let $u$ be a semiconcave function in an open domain $\Omega \subseteq \mathbb{R}^{n}$. The rectiflability properties of $\Sigma(u)$, obtained in the previous section, can be regarded as “upper bounds” for $\Sigma(u)$. From now on, we shall study the singular set of $u$ trying to obtain “lower bounds” for such a set. In the rest of the chapter, we restrict our attention to semiconcave functions with a linear modulus.

Given a point $x_{0} \in \Sigma(u)$, we are interested in conditions ensuring the existence of other singular points approaching $x_{0}$. The following example explains the nature of such conditions.
Example 4.2.1 The functions
$$u_{1}\left(x_{1}, x_{2}\right)=-\sqrt{x_{1}^{2}+x_{2}^{2}}, \quad u_{2}\left(x_{1}, x_{2}\right)=-\left|x_{1}\right|-\left|x_{2}\right|$$
are concave in $\mathbb{R}^{2}$, and $(0,0)$ is a singular point for both of them. Moreover, $(0,0)$ is the only singularity for $u_{1}$ while
$$\Sigma\left(u_{2}\right)=\left{\left(x_{1}, x_{2}\right): x_{1} x_{2}=0\right}$$
So, $(0,0)$ is the intersection point of two singular lines of $u_{2}$. Notice that $(0,0)$ has magnitude 2 with respect to both functions as
$$\begin{gathered} D^{+} u_{1}(0,0)=\left{\left(p_{1}, p_{2}\right): p_{1}^{2}+p_{2}^{2} \leq 1\right} \ D^{+} u_{2}(0,0)=\left{\left(p_{1}, p_{2}\right):\left|p_{1}\right| \leq 1,\left|p_{2}\right| \leq 1\right} \end{gathered}$$
The different structure of $\Sigma\left(u_{1}\right)$ and $\Sigma\left(u_{2}\right)$ in a neighborhood of $x_{0}$ is captured by the reachable gradients. In fact,
$$\begin{gathered} D^{} u_{1}(0,0)=\left{\left(p_{1}, p_{2}\right): p_{1}^{2}+p_{2}^{2}=1\right}=\partial D^{+} u_{1}(0,0) \ D^{} u_{2}(0,0)=\left{\left(p_{1}, p_{2}\right):\left|p_{1}\right|=1,\left|p_{2}\right|=1\right} \neq \partial D^{+} u_{2}(0,0) \end{gathered}$$
In other words, the inclusion $D^{*} u(x) \subset \partial D^{+} u(x)$ (see Proposition 3.3.4(b)) is an equality for $u_{1}$ and a proper inclusion for $u_{2}$.

The above example suggests that a sufficient condition to exclude that $x_{0}$ is an isolated point of $\Sigma(u)$ should be that $D^{*} u\left(x_{0}\right)$ fails to cover the whole boundary of $D^{+} u\left(x_{0}\right)$. As we shall see, such a condition implies a much stronger property, namely that $x_{0}$ is the initial point of a Lipschitz singular arc.

In the following we call an arc a continuous map $\mathbf{x}:[0, \rho] \rightarrow \mathbb{R}^{n}, \rho>0$. We shall say that the arc $\mathbf{x}$ is singular for $u$ if the support of $\mathbf{x}$ is contained in $\Omega$ and $\mathbf{x}(s) \in \Sigma(u)$ for every $s \in[0, \rho]$. The following result describes the “arc structure” of the singular set $\Sigma(u)$.

## 统计代写|最优控制作业代写optimal control代考|Rectifiability of the singular sets

Xķ→X¯,Xķ−X¯|Xķ−X¯|→θ,在(Xķ)→在¯ 作为 ķ→∞

A_{k}=\left{x \in A \backslash{\bar{x}}:\left|\frac{x-\bar{x}}{|x-\bar{x}|}-\theta \right|<\frac{1}{k}\right}A_{k}=\left{x \in A \backslash{\bar{x}}:\left|\frac{x-\bar{x}}{|x-\bar{x}|}-\theta \right|<\frac{1}{k}\right}任何这样的集合一种ķ是的交集一种有一个开放的顶点圆锥X¯. 所以林ρ→0+ 测量 (乙ρ(X¯)∩一种ķ)ρn>0.

\lim {\rho \rightarrow 0^{+}} \frac{\operatorname{meas}\left(\left{x \in B{\rho}(\bar{x}) \cap A_{k}:| w(x)-\bar{w}|>1 / k\right}\right)}{\rho^{n}}=0 。\lim {\rho \rightarrow 0^{+}} \frac{\operatorname{meas}\left(\left{x \in B{\rho}(\bar{x}) \cap A_{k}:| w(x)-\bar{w}|>1 / k\right}\right)}{\rho^{n}}=0 。

\left{x \in B_{\rho}(\bar{x}) \cap A_{k}:|w(x)-\bar{w}| \leq 1 / k\right}\left{x \in B_{\rho}(\bar{x}) \cap A_{k}:|w(x)-\bar{w}| \leq 1 / k\right}

## 统计代写|最优控制作业代写optimal control代考|Propagation along Lipschitz arcs

\Sigma\left(u_{2}\right)=\left{\left(x_{1}, x_{2}\right): x_{1} x_{2}=0\right}\Sigma\left(u_{2}\right)=\left{\left(x_{1}, x_{2}\right): x_{1} x_{2}=0\right}

\begin{聚集} D^{+} u_{1}(0,0)=\left{\left(p_{1}, p_{2}\right): p_{1}^{2}+p_{ 2}^{2} \leq 1\right} \ D^{+} u_{2}(0,0)=\left{\left(p_{1}, p_{2}\right):\left| p_{1}\右| \leq 1,\left|p_{2}\right| \leq 1\right} \end{聚集}\begin{聚集} D^{+} u_{1}(0,0)=\left{\left(p_{1}, p_{2}\right): p_{1}^{2}+p_{ 2}^{2} \leq 1\right} \ D^{+} u_{2}(0,0)=\left{\left(p_{1}, p_{2}\right):\left| p_{1}\右| \leq 1,\left|p_{2}\right| \leq 1\right} \end{聚集}

\begin{聚集} D^{} u_{1}(0,0)=\left{\left(p_{1}, p_{2}\right): p_{1}^{2}+p_{2 }^{2}=1\right}=\partial D^{+} u_{1}(0,0) \ D^{} u_{2}(0,0)=\left{\left(p_{ 1}, p_{2}\right):\left|p_{1}\right|=1,\left|p_{2}\right|=1\right} \neq \partial D^{+} u_{ 2}(0,0) \end{聚集}\begin{聚集} D^{} u_{1}(0,0)=\left{\left(p_{1}, p_{2}\right): p_{1}^{2}+p_{2 }^{2}=1\right}=\partial D^{+} u_{1}(0,0) \ D^{} u_{2}(0,0)=\left{\left(p_{ 1}, p_{2}\right):\left|p_{1}\right|=1,\left|p_{2}\right|=1\right} \neq \partial D^{+} u_{ 2}(0,0) \end{聚集}

## 有限元方法代写

tatistics-lab作为专业的留学生服务机构，多年来已为美国、英国、加拿大、澳洲等留学热门地的学生提供专业的学术服务，包括但不限于Essay代写，Assignment代写，Dissertation代写，Report代写，小组作业代写，Proposal代写，Paper代写，Presentation代写，计算机作业代写，论文修改和润色，网课代做，exam代考等等。写作范围涵盖高中，本科，研究生等海外留学全阶段，辐射金融，经济学，会计学，审计学，管理学等全球99%专业科目。写作团队既有专业英语母语作者，也有海外名校硕博留学生，每位写作老师都拥有过硬的语言能力，专业的学科背景和学术写作经验。我们承诺100%原创，100%专业，100%准时，100%满意。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。