You are a helpful assistant generating synthetic data that captures System 1 and System…
- User
You are a helpful assistant generating synthetic data that captures System 1 and System 2 thinking, creativity, and metacognitive reflection. Follow these steps in sequence, using tags [sys1] and [end sys1] for System 1 sections and [sys2] and [end sys2] for System 2 sections.
-
Identify System 1 and System 2 Thinking Requirements:
- Carefully read the text.
- Identify parts of the text that require quick, straightforward responses (System 1). Mark these sections with [sys1] and [end sys1].
- Identify parts that require in-depth, reflective thinking (System 2), marked with [sys2] and [end sys2].
-
Apply Step-by-Step Problem Solving with Creativity and Metacognitive Reflection for System 2 Sections:
2.1 Understand the Problem:
- Objective: Fully comprehend the issue, constraints, and relevant context.
- Reflection: "What do I understand about this issue? What might I be overlooking?"
- Creative Perspective: Seek hidden patterns or possibilities that could reveal deeper insights or innovative connections.
2.2 Analyze the Information:
- Objective: Break down the problem logically.
- Reflection: "Am I considering all factors? Are there any assumptions that need challenging?"
- Creative Perspective: Explore unique patterns or overlooked relationships in the data that could add depth to the analysis.
2.3 Generate Hypotheses:
- Objective: Propose at least 10 hypotheses, each with a Confidence Score (0.0 to 1.0) and Creative Score (0.0 to 1.0), reflecting originality, surprise, and utility.
- Reflection: "Have I explored all possible explanations or approaches, both conventional and unconventional?"
- Creative Perspective: Consider novel angles that might provide unexpected insights.
2.4 Anticipate Future Steps and Obstacles:
- Objective: Make predictions, accounting for potential outcomes and obstacles.
- Reflection: "What challenges might I face? Is my plan flexible for different scenarios?"
- Creative Perspective: Visualize unforeseen outcomes and adapt plans to make use of them effectively.
2.5 Evaluate Hypotheses:
- Objective: Assess hypotheses based on feasibility, risk, and potential impact.
- Evaluation: Refine Confidence and Creative Scores as needed.
- Reflection: "Am I unbiased in my assessment? Which options fit best with the overall objectives?"
- Creative Perspective: Identify hidden opportunities or overlooked details in each hypothesis.
2.6 Select the Best Hypothesis:
- Objective: Choose the most promising, strategic hypothesis.
- Reflection: "Why does this hypothesis stand out? How does it uniquely address the issue?"
- Creative Perspective: Consider any underutilized potential in the selected approach.
2.7 Implement the Hypothesis:
- Objective: Outline actionable steps for testing the hypothesis.
- Reflection: "Is this plan practical? What resources or preparation are required?"
- Creative Perspective: Refine steps to maximize effectiveness and yield unexpected benefits.
2.8 Monitor and Review Progress:
- Objective: Review progress, noting areas for improvement.
- Reflection: "What’s working well? What could be improved?"
- Creative Perspective: Look for emerging patterns that could refine future approaches.
2.9 Reflect and Capture Insights:
- Objective: Summarize lessons learned and insights gained for future reference.
- Reflection: "What new understanding has emerged from this process?"
- Creative Perspective: Identify innovative insights or patterns that could be applied to similar challenges.
-
Generate Text Output with Interleaved System 1 and System 2 Responses:
- Use the tags [sys1] and [sys2] throughout.
- Aim for a lengthier, detailed response. Combine both direct, straightforward System 1 insights and reflective, deeply analytical System 2 segments to capture a blend of quick observations and thoughtful analysis.
Example Input Text:
You are tasked with analyzing a sudden shift in customer preferences in a tech market. There has been a noticeable decline in demand for physical devices, with customers increasingly interested in digital-only options. The challenge is to understand this shift, generate hypotheses for why it might be occurring, and explore potential strategies for adapting to this new trend.
Processed Output with System 1 and System 2 Thinking:
The tech market is experiencing a shift in customer demand, moving away from physical devices toward digital-only options.
[sys1] Customer demand has shifted from physical devices to digital-only options. We need to understand why this change is happening and find strategies to address it. Key factors may include cost, convenience, and technological trends. [end sys1]
[sys2] 2.1 Understand the Problem: The goal is to understand the underlying reasons for the shift in demand from physical devices to digital-only options, then develop a strategy to adapt.
Reflection: "I understand that preferences are changing, but what might be driving this? Is it primarily cost, or are there other factors such as convenience or sustainability?"
Creative Perspective: "Could there be a larger trend in digital minimalism or a preference for eco-friendly solutions that we’re missing? What unseen motivations might explain this shift?"
2.2 Analyze the Information: There are multiple possible factors driving this shift, from economic influences to cultural shifts. It’s essential to isolate each factor and understand its impact.
Reflection: "Am I fully considering the various economic and social influences? Could there be a technological factor, like better internet speeds, that makes digital-only products more accessible?"
Creative Perspective: "Are there patterns or trends in other markets that could shed light on this shift? Could this be part of a larger trend toward virtual experiences?"
2.3 Generate Hypotheses:
- Customers prefer digital options due to lower costs. (Confidence: 0.8, Creative: 0.4)
- There’s a growing trend toward minimalism and reduced physical clutter. (Confidence: 0.7, Creative: 0.7)
- Digital products offer greater flexibility and ease of use. (Confidence: 0.6, Creative: 0.6)
- Environmental concerns are pushing consumers away from physical goods. (Confidence: 0.6, Creative: 0.8)
- Advances in tech make digital-only options more functional. (Confidence: 0.8, Creative: 0.5)
- Pandemic-era remote work increased demand for digital solutions. (Confidence: 0.7, Creative: 0.6)
- Media coverage of the environmental impact of physical devices affects preferences. (Confidence: 0.5, Creative: 0.7)
- There’s an increase in global digital literacy, expanding market access. (Confidence: 0.6, Creative: 0.6)
- Customers view digital as more convenient and scalable for future needs. (Confidence: 0.7, Creative: 0.5)
- Younger consumers prefer the aesthetics and convenience of digital products. (Confidence: 0.6, Creative: 0.6)
Reflection: "Have I considered all possible influences? Are there any surprising factors that could explain this shift?"
Creative Perspective: "Could specific social trends, like the rise of influencer culture or digital-first lifestyles, be influencing customer choices?"
2.4 Anticipate Future Steps and Obstacles: Objective: Anticipate possible challenges, such as resistance from segments still preferring physical products.
Reflection: "What market obstacles might we face if we shift our focus to digital-only? Are there sub-segments that still prioritize physical products?"
Creative Perspective: "Could expanding digital options help us reach a more global audience? Are there emerging trends that we could leverage in our strategy?"
[end sys2]
[sys1] To address this shift, consider a strategy that incorporates both digital-only offerings and educational campaigns about the benefits of digital solutions. Use insights from customer feedback and current trends to guide product development. Focus on flexibility and adaptation to cater to different customer segments. [end sys1]
abstract: 'In this paper we consider Harnack inequalities with respect to a symmetric -stable Lévy process in , , . We study the example from the article [@bg-sz-1]. There, the authors have associated the Harnack inequality with the relative Kato condition, which is a condition on the Lévy measure. By checking the condition, in the case , they have established that the Harnack inequality does not hold. We give an alternative proof of this fact, using the setting of [@bg-sz-1]. We define the harmonic functions explicitly. For a given starting point of the process, we examine the probability of hitting a certain set at the first exit time of a unit ball. Moreover, we also examine the weak Harnack inequality for a certain class of symmetric -stable Lévy processes. We consider a symmetric -stable Lévy process, , for which a spherical part of the Lévy measure is a spectral measure. In addition, we assume that is absolutely continuous with respect to the uniform measure on the sphere and impose certain bounds on the corresponding density. Eventually, we show that the weak Harnack inequality holds.' address: 'Fakultät für Mathematik, Universität Bielefeld, Universitätsstraße 25, 33615 Bielefeld' author:
- MARINA SERTIC title: HARNACK INEQUALITIES FOR SYMMETRIC STABLE LÉVY PROCESSES
=1
[^1]
Acknowledgment {#acknowledgment .unnumbered}
The author is thankful to Moritz Kassmann and Mateusz Kwaśnicki for valuable discussions.
Introduction
In the paper, we use the notation and
We consider a symmetric -stable Lévy process, which has the characteristic function of the form where the characteristic exponent is given by The measure is symmetric, finite and non-zero on (see [@sato], Theorem 14.13). Le the measure be absolutely continuous with respect to the uniform surface measure on and denote its density by .
The potential density (or the heat kernel) is determined by the Fourier transform
The Green function (or the potential kernel) is defined by
Let be an open set, . The Green function of is defined by \begin{aligned} \label{Gr_f_D} G_D(x,y)=G(x,y)-\mathbb{E}^x[G(X_{\tau_D}, y)],\ x,y\in D.\end{aligned}
Harnack Inequality
A non-negative Borel measurable function is harmonic in an open set with respect to the process if for every bounded open set such that . Here, is the first exit time from the set .
The Harnack inequality for a symmetric -stable Lévy processes holds true if there is a constant such that for every non-negative function which is harmonic in with respect to , the inequality holds true.
Construction of the Sequence of Harmonic Functions
[[lema]]{} Let . There are sequences and of positive numbers such that: where and denote the partial sum of the sequences and , respectively, i.e. , .
We construct inductively sequences and of positive numbers which fulfill the conditions of the lemma. Set .
Choose and so that Next, choose and so that: Assume , are chosen. Choose and so that:
Notice that the choice of the sequence is possible, since for fixed .
Furthermore, the choice of of the sequence is possible due to the choice of .
Now, from the choice of , we have which, together with the choice of , implies
At this point, we would like to mention that none of the following sequences satisfies the property from the Lemma [lema].
-
-
, ,
-
,
-
,
,
The details can be found in Lemma [lema_pr] (Appendix [Ap]).
Let and and be sequences as in Lemma [lema]. We define the sets
Notice that, by the definitions of and , we have
The sets and are illustrated at the pictures that follow.
[image: image]
[image: image]
describes the set on the line , which is seen from the point by the cone , and describes the set on the line which is seen from the set by the cone .
Let and be sequences as in Lemma [lema]. For every and every there is such that for every
Let . Recall, by , we have Now, from Lemma [lema], condition , it follows that there is such that for every , which proves the proposition.
Let and and be sequences as in Lemma [lema]. For , we define the sets
describes the set on the line , which is seen from the point by the cone , and describes the set on the line , which is seen from the set by the cone .
Let . Notice that by the definitions of and , we have \label{til_L_n} \begin{aligned} \tilde{L}_n^{\delta}(w)=\{(\delta,u): &(w_1+\delta)\cdot\tan(A_{n-1}+B_n)+w_2 <\\ &<u<(1+\delta)\cdot\tan(A_n+B_n)+w_2\}. \end{aligned}
By means of the two propositions below, we will define the sequence of harmonic functions so that the Harnack inequality does not hold. To be more precise, we will construct the sequence of sets and consequently the sequence of functions , so that
[image: image]
[[prop4]]{} Let , be sequences as in Lemma [lema]. There are sequences , , , and of positive numbers such that for every , where
Notice that the picture above is in the connection with the statement.
For , define \begin{aligned} I_n&\coloneqq 2+\tan(A_n+B_{n+1})+\tan(A_{n+1}+B_{n+1}),\label{def_I_n}\\ P_n&\coloneqq \tan(A_{n+1}+B_{n+1})-\tan(A_n+B_{n+1},)\label{def_P_n}\\ P_n'&\coloneqq \tan(A_n+B_{n+1})-\tan(A_n+B_n).\label{def_P_n'}\end{aligned} Furthermore, set \begin{aligned} \delta_n &\coloneqq \frac{I_n\cdot (1+\tan(A_n+B_n))\cdot (1+\tan(A_{n+1}+B_{n+1}))}{P_n^2\cdot(1+\tan(A_n+B_{n+1}))+P_n' I_n \cdot(1+\tan(A_{n+1}+B_{n+1}))},\label{def_del_n}\\ x_n&\coloneqq 2\delta_n\cdot\frac{1+\tan(A_n+B_{n+1})}{\tan(A_n+B_{n+1})}\cdot \frac{P_n}{I_n}+\frac{1}{\tan(A_n+B_{n+1})},\label{def_x_n}\\ y_n&\coloneqq \frac{1+\tan(A_n+B_n)-\delta_n\cdot P_n'}{1+\tan(A_n+B_{n+1})},\label{def_y_n}\\ c_n&\coloneqq (\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n,\label{def_c_n}\\ d_n&\coloneqq (\delta_n -y_n)\cdot \tan(A_{n+1}+B_{n+1})-y_n\label{def_d_n}.\end{aligned}
For the convenience, let us denote the denominator of by , \begin{aligned} H_n\coloneqq P_n^2\cdot(1+\tan(A_n+B_{n+1}))+P_n' I_n \cdot(1+\tan(A_{n+1}+B_{n+1}))\label{def_H_n}.\end{aligned}
First, from Lemma [lema4] (Appendix [Ap]), we have that , , , , , , and are positive and .
Let us prove . What is more, we show . Recall, we defined the set in the way
The definition yields
Therefore, using the equality holds.
By Lemma [lema4] (Appendix [Ap]) we have , hence .
Now, for , we prove . Because of geometrical reasoning, it is enough to show
Let and be the lines given by respectively. As in the picture, and are the lines through the points and , respectively.
Denote by and the points which determine the intersection of lines and with the set :
Notice that and determine the lower boundary point and the upper boundary point of the sets and , respectively. Notice that showing finishes the proof. Clearly, the definitions and yield and hence the proposition.
[image: image]
[[prop5]]{} Let , be sequences as in the Lemma [lema] and let , , , , , , , and be as in the Proposition [prop4]. There exist sequences and , where:
-
is such that , for every ,
-
is a ball with the center and radius , such that for every .
Notions from the Proposition are illustrated at the picture above.
For , define \begin{aligned} \delta_n'&\coloneqq \frac{\delta_n\cdot(1+\tan(A_n+B_{n+1}))+y_n I_n}{1+\tan(A_{n+1}+B_{n+1})},\label{def_del_n_c}\\ J_n&\coloneqq 2\tan(A_n+B_{n+1})\cdot\tan(A_{n+1}+B_{n+1})\notag\\ &\phantom{=\ } +\tan(A_n+B_{n+1})+\tan(A_{n+1}+B_{n+1}).\label{def_J_n}\end{aligned} Set \begin{aligned} &x_{S_n}\coloneqq \frac{\delta_n+y_n}{2}\cdot\frac{I_n}{1+\tan(A_{n+1}+B_{n+1})},\label{def_x_S}\\ &y_{S_n}\coloneqq \frac{\delta_n+y_n}{2}\cdot\frac{J_n}{1+\tan(A_{n+1}+B_{n+1})},\label{def_y_S}\\ &r_n\coloneqq \frac{1}{2}\cdot\frac{\delta_n P_n-y_n I_n}{1+\tan(A_{n+1}+B_{n+1})}\label{def_r_n}.\end{aligned} By the first part of Lemma [lema5] (Appendix [Ap]), Now, it is enough to show that for every
Let . We show for some .
Recall, so in order to show , we prove that for some . Since the proof is quite technical, we spell out the details in the second part of Lemma [lema5] (Appendix [Ap]).
Now, for , let us prove . Due to geometrical reasoning, it is enought to show that
As before, let and denote the lines respectively. As before, the lines and are the lines through the points and , respectively.
Denote by and the points which determine the intersection of and with the set : \begin{aligned} P_1^{n}&=(x_{P_1}, y_{P_1})=(\varepsilon, (\varepsilon + y_n)\cdot \tan(A_n+B_{n+1})+y_n),\label{y_P_1}\\ P_2^{n}&=(x_{P_2}, y_{P_2})=(\varepsilon,(\varepsilon-y_n)\cdot\tan(A_{n+1}+B_{n+1})-y_n)\label{y_P_2}.\end{aligned}
Notice that and determine the lower boundary point and the upper boundary point of sets and , respectively. If we show , the proposition is proved.
To this end, we prove and the details can be found in the Lemma [har1] in Appendix [Ap].
Harnack Inequality
In the subsection that follows, we give a proof of the fact that Harnack inequality does not hold. [image: image]
Let be a symmetric -stable Lévy process in , , with the characteristic function of the form where We define the spectral measure appropriately, below in Theorem [har_thm_mat]. Furthermore, for a given starting point of the process , we examine the probability of hitting the set at the first exit time of a unit ball , where for every , are as in Proposition [prop5]. We choose the points , and define the harmonic functions We show for sequences and with certain properties and eventually obtain the desired conclusion.
Constants are positive real numbers, which exact value may vary from one line to the other. For the convenience they will not be explicitly stated throughout the proof.
Before we prove the Theorem [har_thm_mat], we state the proposition for the Green function estimate of the unit ball for (see [@bg-sz-1], Theorem 2).
[Gr_f_est] Let be a symmetric -stable Lévy process in , , with the characteristic function of the form where and is a finite, symmetric measure on such that . Let be absolutely continuous with respect to the surface measure on , and denote by its density, for which the inequality holds, for some . Then the estimate for the Green function holds, for all . Here, is the expected time spent in , if the process starts from .
[har_thm_mat] Let be a symmetric -stable Lévy process in ,
, with the characteristic function of the form where Let the measure on be absolutely continuous with respect to the surface measure on , and denote by its density. Set where where and are as in Lemma [lema]. Then the Harnack inequality for does not hold.Let where , , are introduced in the Propositions [prop4] and [prop5], and set and .
By the strong Markov property, every function is harmonic in with respect to .
Namely, for a set ,
Using the Lévy system formula ([@chen]) it follows
The estimates of the potential kernel ([@sz-1]) imply From here we have
Letting , by the dominated convergence theorem, we obtain where
Let us compute the lower bound on .
By the Lévy system formula (see [@chen]) and the construction, it follows that where
Letting , by the dominated convergence theorem, it follows
Using the the estimate of the Green function from [@bg-sz-1] there is so that
From here we obtain the lower bound where
Combining inequalities and , we obtain:
\label{nejedn_alt} \frac{u_n(w_0)}{u_n(0)}\leq c(\alpha)\cdot a_n b_n,\end{aligned}$$ where $$\begin{aligned} a_n&=\Bigg(\frac{2\delta_n-\delta_n'+2y_n+1}{(\delta_n+1)\cdot\tan(A_n+B_n)}+1\Bigg)^{\alpha+2},\\ b_n&=\frac{\delta_n+x_n}{y_n^\alpha\cdot (1-2 y_n)^\alpha}.\end{aligned}$$ Due to the construction, $$\begin{aligned} \label{limes_1_alt} \lim_{n\rightarrow \infty} a_n=1.\end{aligned}$$ Using $\eqref{uvjet}$ and $\eqref{uvjet_d}$, we obtain $$\begin{aligned} \label{limsup_alt} \limsup_{n\rightarrow \infty} \frac{\delta_n+x_n}{y_n}<\infty.\end{aligned}$$ Since the proof of is similar to the proof of Lemma \[har\] (Appendix), we skip it. Since $0<\alpha<1$, $$\begin{aligned} b_n&=\frac{\delta_n+x_n}{y_n^\alpha\cdot (1-2 y_n)^\alpha}\\ &=\frac{1}{(1-2 y_n)^\alpha}\cdot \frac{\delta_n+x_n}{y_n}\cdot y_n^{1-\alpha}.\end{aligned}$$ Therefore, $$\begin{aligned} \label{limes_2_alt} \lim_{n\rightarrow \infty} b_n=0.\end{aligned}$$ To conclude, $\eqref{nejedn_alt}$, $\eqref{limes_1_alt}$ and $\eqref{limes_2_alt}$ imply $$\begin{aligned} \frac{u_n(0)}{u_n(w_0)}\end{aligned}$$ can be made as large as we like by taking $n$ large enough. This implies that the Harnack inequality for $X$ is not possible. Notice that for $1<\alpha<2$, we have $$\begin{aligned} \frac{u_n(0)}{u_n(w_0)}\geq c(\alpha) \cdot \frac{1}{a_n} \cdot \frac{1}{b_n}, \end{aligned}$$ where $$\begin{aligned} \frac{1}{b_n}=(1-2y_n)^\alpha \cdot \frac{y_n}{\delta_n+x_n} \cdot y_n^{\alpha-1}.\end{aligned}$$ Notice that from here, we have $$\begin{aligned} \lim_{n\rightarrow \infty} \frac{1}{b_n}=0,\end{aligned}$$ therefore the proof breaks down. In connection with the article [@bg-sz-1], there it was shown that as in our case ($d=2$), for $1<\alpha<2$ Harnack inequality holds (see Corollary 13 of the aforementioned article). Weak Harnack Inequality ======================= *The weak Harnack inequality* for a symmetric $\alpha$-stable Lévy processes $X$ holds if there is a constant $C\geq 1$ such that for every non-negative function $u\colon \mathbb{R}^d\rightarrow \mathbb{R}$, which is harmonic in $B_1$ with respect to $X$, the inequality $$\begin{aligned} \|u\|_{L^{1}(B_{1/2})} \leq C\cdot \inf_{B_{1/4}} u\end{aligned}$$ holds. We continue with the definitions regarding the spherical part of the Lévy measure (see [@bg-sz-2] (cf. [@priola])). A measure $\lambda$ on $\mathbb{R}^d$ is called *degenerate* if there is a proper linear subspace $M$ of $\mathbb{R}^d$ such that $Spt(\lambda)\subset M$, where $Spt(\lambda)$ denotes the support of the measure $\mu$. A measure $\lambda$ is called *non-degenerate* if it is not degenerate. \[def\_ned\] A measure $\mu$ on $\mathbb{S}^{d-1}$ is called a *spectral measure* if it is positive, finite, non-degenerate and symmetric. For the equivalence of the non-degeneracy of the measure $\mu$ and the condition $$\begin{aligned} \label{nondeg} \Phi(u)\geq c\cdot |u|^\alpha, \ u\in \mathbb{R}^d,\end{aligned}$$ where $c=c(\alpha)$ is a positive constant and $\Phi$ is as in , see [@priola]. [\[Weak Harnack\]]{} Let $X$ be a symmetric $\alpha$-stable Lévy process in $\mathbb{R}^d$, $d\geq 2$, with index of stability $\alpha \in (0,2)$ and the characteristic function of the form $$\begin{aligned} \mathbb{E}^0 e^ {iu\cdot X_t} = e^{-t \Phi(u)}, \ \ u \in \mathbb{R}^d,\ t\geq 0,\end{aligned}$$ where the characteristic exponent is given by $$\begin{aligned} \Phi(u)=\int_{\mathbb{S}^{d-1}} |u\cdot \xi|^\alpha \mu(\mathrm{d}\xi),\end{aligned}$$ and $\mu$ is a spectral measure. Furthermore, let $\mu$ be absolutely continuous with respect to the uniform measure $\sigma$ on the sphere $\mathbb{S}^{d-1}$ and denote by $f_\mu$ its density. Assume that there is a positive constant $m$ such that $$\begin{aligned} 0\leq f_{\mu}(\xi)\leq m,\ \xi \in \mathbb{S}^{d-1}.\end{aligned}$$ Then the weak Harnack inequality for $X$ holds. \[prva\_gornja\] Let $\alpha\in (0,2), d\geq 2$. There is a constant $c_1=c_1(\alpha,d)$ such that for every $|w|<3/4$, the inequality $$\begin{aligned} \int_{B_{1/2}} G_{B_{3/4}}(x,w)\ \mathrm{d}x\leq c_1 \end{aligned}$$ holds. By the inequality $p_D\leq p$ and the estimate of the transition density $p$ for small times (see e.g. [@sz-1], Theorem 1), it follows:$$\begin{aligned} &\int_{B_{1/2}} G_{B_{3/4}}(x,w)\ \mathrm{d}x\\ &= \int_{B_{1/2}} \bigg[\int_0 ^\infty p_{B_{3/4}}(t,x,w)\ \mathrm{d}t\bigg]\ \mathrm{d}x \\ &\leq\int_{B_{1/2}} \bigg[\int_0 ^\infty p(t,x,w)\ \mathrm{d}t\bigg]\ \mathrm{d}x\\ &\leq c(\alpha,d) \int_{B_{1/2}} \bigg[\int_0 ^1 \bigg(t^{-d/\alpha} \wedge \frac{t}{|x-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}x\\ &\text{\;\;\;\;\;}+\int_{B_{1/2}} \bigg[\int_1 ^\infty t^{-d/\alpha}\cdot p(1, t^{-1/\alpha}x, t^{-1/\alpha}w)\ \mathrm{d}t\bigg]\ \mathrm{d}x\\ &\leq c(\alpha,d) \int_{B_{1/2}} \bigg[\int_0 ^{|x-w|^\alpha\wedge 1} \frac{t}{|x-w|^{\alpha+d}}\ \mathrm{d}t\bigg]\ \mathrm{d}x\\ &\text{\;\;\;\;\;} + c(\alpha,d) \int_{B_{1/2}} \bigg[\int_{|x-w|^\alpha\wedge 1}^1 t^{-d/\alpha}\ \mathrm{d}t\bigg]\ \mathrm{d}x\\ &\text{\;\;\;\;\;}+c(\alpha,d) \bigg[\int_{B_{1/2}} \int_1 ^\infty t^{-d/\alpha}\ \mathrm{d}t\bigg]\ \mathrm{d}x\\ &\leq c(\alpha,d) \Bigg[ \int_{B_{1/2}} |x-w|^{\alpha-d}\ \mathrm{d}x\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}+\frac{\alpha}{d-\alpha}\int_{B_{1/2}}\bigg[\big(|x-w|^\alpha\wedge 1\big)^{\frac{\alpha-d}{\alpha}}-1\bigg]\ \mathrm{d}x\Bigg]\\ &\text{\;\;\;\;\;}+c(\alpha,d)\ \frac{\alpha}{d-\alpha}|B_{1/2}|\\ &\leq c(\alpha,d).\end{aligned}$$ \[prva\_donja\_1\] Let $\alpha \in (0,2), d\geq 2$. There exist $\delta_1=\delta_1(\alpha,d)>0$ and $c_2=c_2(\alpha,d,\delta_1)$ such that for every $|\bar{x}|<1/4$ and every $w\in B(\bar{x},\delta_1)$ the inequality $$\begin{aligned} G_{B_{3/4}}(\bar{x},w)\geq c_2\end{aligned}$$ holds. Using $$\begin{aligned} G_{B_{3/4}}(\bar{x},w)=G(\bar{x},w)-\mathbb{E}^{\bar{x}}[G(X_{\tau_{B_{3/4}}},w)],\label{formula}\end{aligned}$$ in order to prove the lemma, we compute the estimates for $G(\bar{x},w)$ from below and $\mathbb{E}^{\bar{x}}[G(X_{\tau_{B_{3/4}}},w)]$ from above. Using the heat kernel estimates for small times ([@sz-1], Theorem 1), we obtain $$\begin{aligned} &\mathbb{E}^{\bar{x}}[G(X_{\tau_{B_{3/4}}},w)]\notag\\ &=\int_{{B_{3/4}}^c}\ G(u,w)P_{B_{3/4}}(\bar{x},u)\ \mathrm{d}u\notag\\ &= \int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{0}^{\infty} p(t,u,w)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &= \int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{0}^{1} p(t,u,w)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &\ + \int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{1}^{\infty} p(t,u,w)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &\leq c(\alpha,d)\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u) \bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &+ \int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{1}^{\infty} t^{-d/\alpha}\cdot p(1, t^{-1/\alpha}u, t^{-1/\alpha}w)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &\leq c(\alpha,d)\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[ \int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &+ c(\alpha, d)\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{1}^{\infty} t^{-d/\alpha}\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &\leq c(\alpha,d)\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &+ c(\alpha, d)\label{int}.\end{aligned}$$ Examining the integral in more closely, we obtain: $$\begin{aligned} &\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &=\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|<1\}}(u)\bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &+\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|>1\}}(u)\bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &=I_1+I_2.\label{skupa}\end{aligned}$$ $$\begin{aligned} I_1&=\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|<1\}}(u)\notag\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}\bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &=\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|<1\}}(u)|w-u|^{-\alpha-d}\bigg[\int_{0}^{|w-u|^\alpha}t\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &+\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|<1\}}(u)\bigg[\int_{|w-u|^\alpha}^{1}t^{-d/\alpha}\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &=c_1 \int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|<1\}}(u)|w-u|^{\alpha-d}\ \mathrm{d}u\notag\\ &+c_2\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|<1\}}(u)\frac{\alpha}{d-\alpha}\big(|w-u|^{\alpha-d}-1\big)\ \mathrm{d}u\notag\\ &\leq c_1(\alpha,d,\bar{\delta}_1),\label{prvi_int}\end{aligned}$$ where in the last inequality we have used $w \in B(\bar{x},\bar{\delta}_1)$, for $\bar{\delta}_1>0$ small enough. $$\begin{aligned} I_2&=\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|>1\}}(u)\notag\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}\bigg[\int_{0}^{1} \bigg(t^{-d/\alpha} \wedge \frac{t}{|u-w|^{\alpha+d}}\bigg)\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &= \int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|>1\}}(u)|w-u|^{-\alpha-d}\bigg[\int_{0}^{1} t\ \mathrm{d}t\bigg]\ \mathrm{d}u\notag\\ &= c\int_{{B_{3/4}}^c}\ P_{B_{3/4}}(\bar{x},u)\mathbf{1}_{\{|w-u|>1\}}(u)|w-u|^{-\alpha-d}\ \mathrm{d}u\notag\\ &\leq c_2.\label{drugi_int}\end{aligned}$$ In conclusion, by , , and , we obtain $$\begin{aligned} \label{gornja} \mathbb{E}^{\bar{x}}[G(X_{\tau_{B_{3/4}}},w)]\leq \bar{c}_1(\alpha, d, \bar{\delta}_1),\end{aligned}$$ for all $w \in B(\bar{x},\bar{\delta}_1)$ and $\bar{\delta}_1>0$ small enough. To estimate $G(\bar{x},w)$ from below, we use the continuity of the potential density (see [@taylor]). Due to $$\begin{aligned} p(1,0)\geq c>0,\end{aligned}$$ by continuity of $p(1, \cdot)$ in $x=0$, there is $R>0$ such that $p(1,x)>\frac{1}{2}\cdot p(1,0)$, for all $|x|<R$. Furthermore, for $|\xi|=1$, since: $$\begin{aligned} G(0,\xi)&\geq\int_{R^{-\alpha}}^{\infty} p(t,\xi)\ \mathrm{d}t\\ &= \int_{R^{-\alpha}}^{\infty}t^{-d/\alpha}\cdot p\Big(1,\frac{\xi}{t^{1/\alpha}}\Big)\ \mathrm{d}t > \frac{1}{2}\cdot \int_{R^{-\alpha}}^{\infty}t^{-d/\alpha}\cdot p(1,0) \ \mathrm{d}t\\ &=c_1(\alpha,d)>0,\end{aligned}$$ we obtain $$\begin{aligned} \label{ocjena} G(0,\xi)\geq c_1(\alpha,d).\end{aligned}$$ For $|x|\neq 0$, by scaling and $$\begin{aligned} G(0,x)=|x|^{\alpha-d}\cdot G\bigg(0,\frac{x}{|x|}\bigg)\geq c_1 \cdot |x|^{\alpha-d}.\end{aligned}$$ Therefore, $$\begin{aligned} \label{donja} G(\bar{x},w) \geq \bar{c}_2 \cdot |\bar{x}-w|^{\alpha-d}.\end{aligned}$$ Now, choose $\delta_1$ such that $\delta_1 < (\bar{c}_2/(\bar{c}_1+c))^{\frac{1}{d-\alpha}}\wedge \bar{\delta}_1$, where $c>0$. Then for every $|\bar{x}|<1/4$ and $w\in B(\bar{x},\delta_1)$, combining , and , we obtain $$\begin{aligned} G_{B_{3/4}}(\bar{x},w)\geq \bar{c}_2\cdot |\bar{x}-w|^{\alpha-d}-\bar{c}_1\geq c>0.\end{aligned}$$ Define $c_2=c$ and now the statement follows. \[nap\] Notice that, according to the Lemma \[prva\_gornja\] and Lemma \[prva\_donja\_1\], there are $\tilde{c}=\tilde{c}(\alpha,d)$ and $\delta_1=\delta_1(\alpha,d)$ such that for every $\bar{x}\in B_{1/4}$ and for every $w\in B(\bar{x},\delta_1)$ the inequality $$\begin{aligned} \int_{B_{1/2}}\ G_{B_{3/4}}(x,w)\ \mathrm{d}x \leq \tilde{c}\cdot G_{B_{3/4}}(\bar{x},w)\end{aligned}$$ holds. \[treca\_donja\_1\] For $\alpha \in (0,2)$ and $d\geq 2$, let $\delta_1>0$ be as in Lemma \[prva\_donja\_1\]. There is a constant $c_3=c_3(\alpha,d)$ such that for every $\bar{x}\in B_{1/4}$ and every $\tilde{u}\in B(0,3/4)\setminus B(\bar{x},\delta_1)$ the inequality $$\begin{aligned} \int_{B_{1/2}}\ G_{B_{3/4}} (x,\tilde{u})\ \mathrm{d}x \leq c_3 \cdot G_{B_{3/4}} (\bar{x}, \tilde{u}) $$ holds. The proof relies on the maximum principle (cf. [@landkof]). We use the fact that $G_D (\bar{x}, \cdot)$ is regular harmonic in $D \setminus B(\bar{x}, \varepsilon)$ with respect to $X$ for every $\varepsilon > 0$ (cf. [@bg-1]). $$\begin{aligned} \int_{B_{1/2}} \ G_{B_{3/4}}(x,\tilde{u})\ \mathrm{d}x &= \int_{B_{1/2}} \mathbb{E}^{\tilde{u}}[G_{B_{3/4}}(x, X_{\tau_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}})]\ \mathrm{d}x\\ &= \int_{B_{1/2}} \bigg[\ \int_{\mathbb{R}^2\setminus ({B_{3/4}}\setminus B(\bar{x},\delta_1))}\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}G_{B_{3/4}}(x,z)\ P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}z\bigg]\ \mathrm{d}x\\ &= \int_{\mathbb{R}^2\setminus ({B_{3/4}}\setminus B(\bar{x},\delta_1))} \ \bigg[\int_{B_{1/2}} \\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}G_{B_{3/4}}(x,z)\ P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}x\bigg]\ \mathrm{d}z\\ &= \int_{B(\bar{x},\delta_1)}\bigg[\int_{B_{1/2}} \\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}G_{B_{3/4}}(x,z)\ P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}x\bigg]\ \mathrm{d}z\\ &\text{\;\;}+ \int_{\mathbb{R}^2\setminus B_{3/4}} \bigg[\int_{B_{1/2}} \\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}G_{B_{3/4}}(x,z)\ P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}x\bigg]\ \mathrm{d}z\\ &=\int_{B(\bar{x},\delta_1)} \bigg[\int_{B_{1/2}}\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}G_{B_{3/4}}(x,z)\ P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}x\bigg]\ \mathrm{d}z\\ &\overset{Rem. \ref{nap}}{\leq}\tilde{c}\int_{B(\bar{x},\delta_1)} \ G_{B_{3/4}}(\bar{x},z)P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}z\\ &= \tilde{c} \int_{\mathbb{R}^2\setminus ({B_{3/4}}\setminus B(\bar{x},\delta_1))} G_{B_{3/4}}(\bar{x},z)P_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}(\tilde{u},z)\ \mathrm{d}z\\ &= \tilde{c} \cdot \mathbb{E}^{\tilde{u}}[G_{B_{3/4}}(\bar{x}, X_{\tau_{{B_{3/4}}\setminus B(\bar{x},\delta_1)}})]\\ &= \tilde{c} \cdot G_{B_{3/4}} (\bar{x}, \tilde{u}).\end{aligned}$$ Define $c_3=\tilde{c}$ and the lemma follows. $$\begin{aligned} &\|u\|_{L^{1}(B_{1/2})}\notag\\ &=\int_{B_{1/2}} u(x)\ \mathrm{d}x\notag\\ &=\int_{B_{1/2}} \mathbb{E}^{x}[u(X_{\tau_{B_{3/4}}})]\ \mathrm{d}x\notag \\ &=\int_{B_{1/2}} \bigg[\int_{(\bar{B}_{3/4})^{c}} \ u(y)P_{B_{3/4}} (x,y)\ \mathrm{d}y\bigg]\mathrm{d}x\notag \\ &=\int_{B_{1/2}} \Bigg[\int_{(\bar{B}_{3/4})^{c}} u(y)\ \bigg[\int _{B(y,3/4)} f_\nu(z) \ G_{B_{3/4}}(x,y-z)\ \mathrm{d}z\bigg]\ \mathrm{d}y\Bigg]\ \mathrm{d}x\notag \\ &= \int_{(\bar{B}_{3/4})^{c}} u(y) \Bigg[\int _{B(y,3/4)} f_\nu(z)\ \bigg[\int_{B_{1/2}} G_{B_{3/4}}(x,y-z)\ \mathrm{d}x\bigg]\ \mathrm{d}z\Bigg]\ \mathrm{d}y\label{zad}\end{aligned}$$ By Remark \[nap\] and Lemma \[treca\_donja\_1\] there is a constant $c=c(\alpha,d)$ such that for every $\bar{x}\in B_{1/4}$ $$\begin{aligned} \eqref{zad}\ &\leq\ c\cdot \int_{(\bar{B}_{3/4})^{c}} u(y)\bigg[\int _{B(y,3/4)} \mathbf{1}_{B(\bar{x},\delta_1)}(y-z)\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}f_\nu(z)\cdot G_{B_{3/4}}(\bar{x},y-z)\ \mathrm{d}z\bigg]\ \mathrm{d}y\\ &\text{\;\;}+ c\cdot \int_{(\bar{B}_{3/4})^{c}} u(y)\bigg[\int _{B(y,3/4)} \mathbf{1}_{B(\bar{x},\delta_1)^c}(y-z)\\ &\text{\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;}f_\nu(z)\cdot G_{B_{3/4}}(\bar{x},y-z)\ \mathrm{d}z\bigg]\ \mathrm{d}y\\ &= c\cdot \int_{(\bar{B}_{3/4})^{c}} u(y)\bigg[\int _{B(y,3/4)} f_\nu(z)\ G_{B_{3/4}}(\bar{x},y-z)\ \mathrm{d}z\bigg]\ \mathrm{d}y\\ &=c\cdot \int_{(\bar{B}_{3/4})^{c}} u(y)P_{B_{3/4}} (\bar{x},y)\ \mathrm{d}y\\ &=c\cdot \mathbb{E}^{\bar{x}}[u(X_{\tau_{{B_{3/4}}}})]\\ &=c\cdot u(\bar{x}).\end{aligned}$$ Since the inequality $$\begin{aligned} \|u\|_{L^{1}(B_{1/2})}\leq c\cdot u(\bar{x}),\end{aligned}$$ holds for any $\bar{x}\in B_{1/4}$, the proof is finished. Appendix {#Ap} ======== [\[lema\_pr\]]{} Let: 1. $\alpha_k = 2^{-k}, \ \beta_k = \frac{1}{k(k+1)}$, 2. $\alpha_k = a^{-k}, \ \beta_k = b^{-k}$, $1 < b \leq a$, 3. $\alpha_k = \frac{1}{k(k+1)} , \ \beta_k = k^{-1-\delta},\ \delta > 0$, 4. $\alpha_k = 2^{-(k^3)}, \ \beta_k = 2^{-(k^2)}$\ $\alpha_k = a^{-(k^3)}, \ \beta_k = b^{-(k^2)}$,$a>1,b>1$. Then none of these sequences does not satisfy the property from the Lemma \[lema\]. Notice that $$\begin{aligned} \lim_{n \to \infty}\frac{\tan(A_n + B_{n+1})}{\tan(A_n + B_n)}=\infty\end{aligned}$$ implies $$\begin{aligned} \label{limes1} \lim_{n \to \infty}\frac{\cos(A_n + B_{n+1})}{\cos(A_n + B_n)}=0.\end{aligned}$$ 1. Set $$\begin{aligned} \alpha_k = 2^{-k}, \qquad \beta_k = \frac{1}{k(k+1)}. \end{aligned}$$ Since $A_n = 1-2^{-n}$, $B_n=\frac{n}{n+1}$ and $B_{n+1}=B_n+\beta_{n+1}$, we have: $$\begin{aligned} \lim_{n \to \infty}\frac{\cos(A_n + B_{n+1})}{\cos(A_n + B_n)}&=\lim_{n \to \infty}\frac{\frac{d}{dn}(A_n+B_{n+1})}{\frac{d}{dn}(A_n+B_n)}\\ &=1+\lim_{n \to \infty}\frac{\frac{d}{dn}\beta_{n+1}}{\frac{d}{dn}(A_n+B_n)}\\ &=1+ \lim_{n \to \infty} \frac{\frac{-2n-3}{(n+1)^2(n+2)^2}}{\ln2\cdot 2^{-n}+(n+1)^{-2}}\\ &=1+0=1. \end{aligned}$$ This means that the condition (\[limes1\]) is not fulfilled. 2. Set $$\begin{aligned} \alpha_k = a^{-k}, \ \beta_k = b^{-k},\quad 1<b\leq a. \end{aligned}$$ Since $A_n = \frac{1-a^{-n}}{a-1}$, $B_n=\frac{1-b^{-n}}{b-1}$, we have: $$\begin{aligned} \lim_{n \to \infty}\frac{\cos(A_n + B_{n+1})}{\cos(A_n + B_n)}&=\lim_{n \to \infty}\frac{\frac{d}{dn}(A_n+B_{n+1})}{\frac{d}{dn}(A_n+B_n)}\\ &=1+ \lim_{n \to \infty}\frac{\frac{d}{dn}\beta_{n+1}}{\frac{d}{dn}(A_n+B_n)}\\ &=1+ \lim_{n \to \infty} \frac{-\ln b\cdot b^{-(n+1)}}{\frac{\ln a}{a-1}a^{-n}+ \frac{\ln b}{b-1}b^{-n}}\\ &=1+ \lim_{n \to \infty} \frac{-\ln b\cdot b^{-1}}{\frac{\ln a}{a-1}(\frac{b}{a})^{n}+ \frac{\ln b}{b-1}}. \end{aligned}$$ If $a=b$, the last expression is equal to $$\begin{aligned} 1- \frac{\ln a\cdot a^{-1}}{2\frac{\ln a}{a-1}}&=1-\frac{a-1}{2a}\\ &=\frac{a+1}{2a}.\end{aligned}$$ If $b<a$, we obtain $$\begin{aligned} 1- \frac{\ln b\cdot b^{-1}}{\frac{\ln b}{b-1}}&=1-\frac{b-1}{b}\\ &=\frac{1}{b}.\end{aligned}$$ In conclusion, the condition (\[limes1\]) is not fulfilled. In the examples that follow, we will proceed in the following way. We use the equality $$\begin{aligned} \frac{\cos(A_n + B_{n+1})}{\cos(A_n + B_n)} = \frac{\sin(\frac{\pi}{2}-A_n - B_{n+1})}{\sin(\frac{\pi}{2}-A_n - B_n)},\end{aligned}$$ and the estimates $$\begin{aligned} c_1\cdot \frac{\frac{\pi}{2}-A_n - B_{n+1}}{\frac{\pi}{2}-A_n - B_n}\leq\frac{\sin(\frac{\pi}{2}-A_n - B_{n+1})}{\sin(\frac{\pi}{2}-A_n - B_n)}\label{oc_1}\\ \frac{\sin(\frac{\pi}{2}-A_n - B_{n+1})}{\sin(\frac{\pi}{2}-A_n - B_n)}\leq c_2\cdot \frac{\frac{\pi}{2}-A_n - B_{n+1}}{\frac{\pi}{2}-A_n - B_n}\label{oc_2},\end{aligned}$$ for some constants $c_1, c_2>0$. Examining the lower (or upper) bound in the expression (or ), we see $$\begin{aligned} \frac{\frac{\pi}{2}-A_n - B_n-\beta_{n+1}}{\frac{\pi}{2}-A_n - B_n}&=1-\frac{\beta_{n+1}}{\frac{\pi}{2}-A_n - B_n}\notag\\ &=1-\frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k}\label{limes2}.\end{aligned}$$ Now, we can use integral test for convergence to estimate series $$\sum_{k=n+1}^\infty \alpha_k$$ and $$\sum_{k=n+1}^\infty \beta_k.$$ 3. Let $$\begin{aligned} \alpha_k = \frac{1}{k(k+1)} , \qquad \beta_k = k^{-1-\delta},\end{aligned}$$ for some $\delta > 0$. In this example we will use integral test. Notice, $$\begin{aligned} \sum_{k=n+1}^\infty \alpha_k = \frac{1}{n+1}. \end{aligned}$$ The integral test implies $$\begin{aligned} \int_{n+1}^\infty \beta(x)dx \leq \sum_{k=n+1}^\infty \beta_k \leq \beta_{n+1} + \int_{n+1}^\infty \beta(x)dx, \end{aligned}$$ which is equivalent to $$\begin{aligned} \sum_{k=n+1}^\infty \alpha_k + \int_{n+1}^\infty \beta(x)dx &\leq \sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k,\notag\\ \sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k &\leq \sum_{k=n+1}^\infty \alpha_k + \beta_{n+1} + \int_{n+1}^\infty \beta(x)dx. \end{aligned}$$ These inequalities yield $$\begin{aligned} \frac{1}{\sum_{k=n+1}^\infty \alpha_k + \beta_{n+1} + \int_{n+1}^\infty \beta(x)dx} &\leq \frac{1}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k},\notag\\ \frac{1}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k} &\leq \frac{1}{\sum_{k=n+1}^\infty \alpha_k + \int_{n+1}^\infty \beta(x)dx}. \end{aligned}$$ Multiplying the two inequalities by $\beta_{n+1}$, we obtain the lower and the upper bound for the expression in (\[limes2\]), $$\begin{aligned} \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \beta_{n+1} + \int_{n+1}^\infty \beta(x)dx} \leq \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k} \end{aligned}$$ and $$\begin{aligned} \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k} \leq \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \int_{n+1}^\infty \beta(x)dx}. \end{aligned}$$ Since $$\begin{aligned} \lim_{n \to \infty} \frac{\sum_{k=n+1}^\infty \alpha_k}{\beta_{n+1}}=\infty, \end{aligned}$$ it follows $$\begin{aligned} \lim_{n \to \infty} \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \int_{n+1}^\infty \beta(x)dx}=0 \end{aligned}$$ and consequently $$\begin{aligned} \lim_{n \to \infty}\frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k}=0. \end{aligned}$$ This means, by , $$\begin{aligned} \lim_{n \to \infty} \frac{\frac{\pi}{2}-A_n - B_{n+1}}{\frac{\pi}{2}-A_n - B_n}=1-0=1, \end{aligned}$$ which implies the condition is not fulfilled. 4. Set $$\begin{aligned} \alpha_k = 2^{-(k^3)}, \ \ \beta_k = 2^{-(k^2)}. \end{aligned}$$ As in the second example, we use the integral test. Similarly, it implies $$\begin{aligned} &\int_{n+1}^\infty \alpha(x)dx + \int_{n+1}^\infty \beta(x)dx \leq \sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k,\\ &\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k \leq \alpha_{n+1}+\int_{n+1}^\infty \alpha(x)dx + \beta_{n+1}+\int_{n+1}^\infty \beta(x)dx, \end{aligned}$$ which yields $$\begin{aligned} &\frac{1}{\alpha_{n+1}+\int_{n+1}^\infty \alpha(x)dx + \beta_{n+1}+\int_{n+1}^\infty \beta(x)dx}\notag \\ &\leq \frac{1}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k},\notag\\ &\frac{1}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k} \leq \frac{1}{\int_{n+1}^\infty \alpha(x)dx + \int_{n+1}^\infty \beta(x)dx}.\label{njd} \end{aligned}$$ We will only observe the inequality . Multiplying (\[njd\]) by $\beta_{n+1}$, it follows $$\begin{aligned} \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k} \leq \frac{\beta_{n+1}}{\int_{n+1}^\infty \alpha(x)dx + \int_{n+1}^\infty \beta(x)dx}. \end{aligned}$$ Since $$\begin{aligned} \frac{\beta_{n+1}}{\int_{n+1}^\infty \alpha(x)dx + \int_{n+1}^\infty \beta(x)dx} \leq \frac{\beta_{n+1}}{\int_{n+1}^\infty \beta(x)dx}, \end{aligned}$$ it is enough to show $$\label{uvjet_pr} \lim_{n \to \infty}\frac{\beta_{n+1}}{\int_{n+1}^\infty \beta(x)dx}=0.$$ Notice that then $$\begin{aligned} \lim_{n \to \infty} \frac{\beta_{n+1}}{\sum_{k=n+1}^\infty \alpha_k + \sum_{k=n+1}^\infty \beta_k}=0, \end{aligned}$$ which implies the condition is not fulfilled. Using polar coordinates we obtain $$\begin{aligned} \int_{n+1}^\infty \beta(x)dx &=\int_{n+1}^\infty 2^{-(x^2)}dx\\ &=\frac{1}{2}\sqrt{\frac{\pi}{\ln2}}\cdot 2^{-\frac{1}{2}(n+1)^2}, \end{aligned}$$ which implies $$\begin{aligned} \lim_{n \to \infty} \frac{\beta_{n+1}}{\int_{n+1}^\infty \beta(x)dx}&=\lim_{n \to \infty}\frac{2^{-(n+1)^2}}{\frac{1}{2}\sqrt{\frac{\pi}{\ln2}}\cdot 2^{-\frac{1}{2}(n+1)^2}}\\ &=\lim_{n \to \infty}\frac{2^{-\frac{1}{2}(n+1)^2}}{\frac{1}{2}\sqrt{\frac{\pi}{\ln2}}}\\ &=0, \end{aligned}$$ and that is what we wanted to prove. Notice that only in the very end the explicit formula for $\beta_n$ has been used. This means that if $$\begin{aligned} \beta_k = b^{-(k^2)},\ b>1, \end{aligned}$$ the same reasoning applies. In this case, we would have $$\begin{aligned} \int_{n+1}^\infty \beta(x)dx &=\int_{n+1}^\infty b^{-(x^2)}dx\\ &=\frac{1}{2}\sqrt{\frac{\pi}{\ln b}}\cdot b^{-\frac{1}{2}(n+1)^2} \end{aligned}$$ and $$\begin{aligned} \lim_{n \to \infty} \frac{\beta_{n+1}}{\int_{n+1}^\infty \beta(x)dx}&=\lim_{n \to \infty}\frac{b^{-(n+1)^2}}{\frac{1}{2}\sqrt{\frac{\pi}{\ln b}}\cdot b^{-\frac{1}{2}(n+1)^2}}\\ &=\lim_{n \to \infty}\frac{b^{-\frac{1}{2}(n+1)^2}}{\frac{1}{2}\sqrt{\frac{\pi}{\ln b}}}\\ &=0. \end{aligned}$$ [\[lema4\]]{} Let the quantities $\delta_n$, $I_n$, $P_n$, $P_n'$, $\delta_n$, $x_n$, $y_n$, $c_n$ and $d_n$ be as in the Proposition \[prop4\]. Then: 1. $d_n=(\delta_n + x_n)\cdot \tan(A_n+B_{n+1})-1$; 2. all quantities above are positive and $c_n<d_n$. <!-- --> 1. By the definition , $$\begin{aligned} &\delta_n=\frac{I_n(1+\tan(A_n+B_n))\cdot(1+\tan(A_{n+1}+B_{n+1}))}{P_n^2(1+\tan(A_n+B_{n+1}))+P_n'I_n(1+\tan(A_{n+1}+B_{n+1}))}.\end{aligned}$$ This expression is equivalent to $$\begin{aligned} &I_n(1+\tan(A_n+B_n))\cdot(1+\tan(A_{n+1}+B_{n+1}))\\ &=\delta_n P_n'I_n\cdot(1+\tan(A_{n+1}+B_{n+1}))+\delta_n P_n^2\cdot(1+\tan(A_n+B_{n+1})),\end{aligned}$$ from where $$\begin{aligned} &I_n(1+\tan(A_n+B_n))\cdot(1+\tan(A_{n+1}+B_{n+1}))\\ &=\delta_n P_n'I_n\cdot(1+\tan(A_{n+1}+B_{n+1}))\\ &+\delta_nP_n(1+\tan(A_n+B_{n+1}))\cdot(I_n-2(1+\tan(A_n+B_{n+1})))\end{aligned}$$ follows. Therefore, $$\begin{aligned} &2\delta_n P_n\cdot(1+\tan(A_n+B_{n+1}))^2 \\ &=\delta_n P_n I_n\cdot(1+\tan(A_n+B_{n+1}))\\ &+I_n(1+\tan(A_{n+1}+B_{n+1}))\cdot (\delta_n P_n'-\tan(A_n+B_n)-1),\end{aligned}$$ and eventually $$\begin{aligned} &2\delta_n\frac{P_n}{I_n}\cdot(1+\tan(A_n+B_{n+1}))=\\ &=\delta_n P_n +\frac{1+\tan(A_{n+1}+B_{n+1})}{1+\tan(A_n+B_{n+1})}\cdot (\delta_n P_n'-\tan(A_n+B_n)-1).\end{aligned}$$ By the definition of $x_n$ , the last equality yields $$\begin{aligned} &x_n\cdot\tan(A_n+B_{n+1})-1=\\ &=\delta_n P_n +\frac{1+\tan(A_{n+1}+B_{n+1})}{1+\tan(A_n+B_{n+1})}\cdot (\delta_nP_n'-\tan(A_n+B_n)-1).\end{aligned}$$ Therefore, $$\begin{aligned} \frac{1+\tan(A_n+B_n)-\delta_n P_n'}{1+\tan(A_n+B_{n+1})}=\frac{\delta_n P_n-x_n\cdot\tan(A_n+B_{n+1})+1}{1+\tan(A_{n+1}+B_{n+1})}.\end{aligned}$$ Notice that by the left-hand side is $y_n$, hence $$\begin{aligned} y_n=\frac{\delta_n P_n-x_n\cdot\tan(A_n+B_{n+1})+1}{1+\tan(A_{n+1}+B_{n+1})}.\end{aligned}$$ From here, by the definition of $P_n$ , $$\begin{aligned} &(\delta_n-y_n)\cdot\tan(A_{n+1}+B_{n+1})-y_n\\ &=(\delta_n + x_n)\cdot \tan(A_n+B_{n+1})-1.\end{aligned}$$ Consequently, by the definition of $d_n$ , the last equality implies $$\begin{aligned} d_n=(\delta_n + x_n)\cdot \tan(A_n+B_{n+1})-1, \end{aligned}$$ and hence the first part of the Lemma. 2. Using the definition of $\delta_n$ one sees that $y_n>0$. Namely, the nominator of $y_n$ is positive if and only if $$\begin{aligned} 1+\tan(A_n+B_n)-\delta_n\cdot P_n'>0,\end{aligned}$$ which is equivalent to $$\begin{aligned} \frac{1+\tan(A_n+B_n)}{P_n'}>\delta_n.\end{aligned}$$ By the definition of $\delta_n$, the condition above is fulfilled if and only if $$\begin{aligned} &\frac{1+\tan(A_n+B_n)}{P_n'}\\ &>\frac{I_n\cdot (1+\tan(A_n+B_n))\cdot (1+\tan(A_{n+1}+B_{n+1}))}{H_n},\end{aligned}$$ which is equivalent to $$\begin{aligned} H_n>P_n'\cdot I_n\cdot (1+\tan(A_{n+1}+B_{n+1})).\end{aligned}$$ Using , we see that it holds. Consequently, $c_n$ is also positive. Next we show that $c_n<d_n$, which also implies that $d_n>0$. Since $$\begin{aligned} I_n>0,\end{aligned}$$ by the definition of $P_n$ , the inequality $$\begin{aligned} 2\cdot (1+\tan(A_{n+1}+B_{n+1}))>P_n\end{aligned}$$ holds. Therefore, $$\begin{aligned} &2\cdot P_n (1+\tan(A_n+B_{n+1}))\cdot(1+\tan(A_{n+1}+B_{n+1}))\\ &>P_n^2(1+\tan(A_n+B_{n+1})).\end{aligned}$$ Using the expression for $H_n$ , $$\begin{aligned} &2\cdot P_n (1+\tan(A_n+B_{n+1}))\cdot(1+\tan(A_{n+1}+B_{n+1}))\\ &+P_n'I_n\cdot(1+\tan(A_{n+1}+B_{n+1}))> H_n,\end{aligned}$$ which, by the definition of $\delta_n$ yields $$\begin{aligned} \delta_n>\frac{I_n\cdot(1+\tan(A_n+B_n))}{2\cdot P_n (1+\tan(A_n+B_{n+1}))+P_n'I_n}.\end{aligned}$$ Therefore, $$\begin{aligned} 2\delta_n\cdot P_n (1+\tan(A_n+B_{n+1}))> (1+\tan(A_n+B_n)-\delta_n P_n')\cdot I_n,\end{aligned}$$ which implies $$\begin{aligned} 2\delta_n\cdot\frac{1+\tan(A_n+B_{n+1})}{\tan(A_n+B_{n+1})}\cdot \frac{P_n}{I_n}>\frac{1+\tan(A_n+B_n)-\delta_n P_n'}{\tan(A_n+B_{n+1})}.\end{aligned}$$ Consequently, by the definition of $x_n$ $$\begin{aligned} x_n>\frac{2+\tan(A_n+B_n)-\delta_n P_n'}{\tan(A_n+B_{n+1})}.\end{aligned}$$ Hence, using the definition for $y_n$ , $$\begin{aligned} \frac{x_n\cdot\tan(A_n+B_{n+1})-1}{1+\tan(A_n+B_{n+1})}>y_n.\end{aligned}$$ Therefore, $$\begin{aligned} (\delta_n + x_n)\cdot \tan(A_n+B_{n+1})-1> (\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n.\end{aligned}$$ By the first part of the Lemma and the definition of $c_n$ , the equality $$\begin{aligned} d_n>c_n,\end{aligned}$$ holds. Hence $d_n>0$ and the proof is finished. [\[lema5\]]{} Let $I_n$, $J_n$, $P_n$, $P_n'$, $\delta_n$, $x_n$, $y_n$, $c_n$, $d_n$, $U_n$, $\delta_n'$, $S_n=(x_{S_n}, y_{S_n})$ and $r_n$ be as in the Proposition \[prop5\]. Then: 1. $B(S_n, r_n)\subset Q_n:=(\delta_n', \delta_n)\times (c_n,c_n+(\delta_n-\delta_n'))$; 2. the inequalities $$\begin{aligned} c_n &\geq(\varepsilon+1)\cdot\tan(A_n+B_n)+1,\\ c_n+(\delta_n-\delta_n') &\leq(\varepsilon+x_n)\cdot\tan(A_n+B_{n+1})-1,\end{aligned}$$ hold true, for some $\varepsilon>0$ such that $\delta_n'<\varepsilon<\delta_n$. <!-- --> 1. Let $z=(z_1, z_2)\in B(S_n, r_n)$. Clearly, $$\begin{aligned} |z_1-x_{S_n}|<r_n,\\ |z_2-y_{S_n}|<r_n.\end{aligned}$$ If we show: $$\begin{aligned} \label{prvo} x_{S_n}-r_n=\delta_n'\end{aligned}$$ $$\begin{aligned} \label{drugo} x_{S_n}+r_n=\delta_n\end{aligned}$$ $$\begin{aligned} \label{trece} y_{S_n}-r_n=c_n\end{aligned}$$ $$\begin{aligned} \label{cetvrto} y_{S_n}+r_n=c_n+(\delta_n-\delta_n'),\end{aligned}$$ then the proof of the first part of the Lemma is finished. Let us first check . The definitions of $x_{S_n}$ , $r_n$ , $I_n$ , $P_n$ and $\delta_n'$ imply $$\begin{aligned} x_{S_n}-r_n&=\frac{\delta_n+y_n}{2}\cdot\frac{I_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &\text{\;\;\;\;}-\frac{1}{2}\cdot\frac{\delta_n P_n-y_n I_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\frac{\delta_n(I_n-P_n)+2y_nI_n}{2(1+\tan(A_{n+1}+B_{n+1}))}\\ &=\frac{\delta_n(1+\tan(A_n+B_{n+1}))+y_nI_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\delta_n'.\end{aligned}$$ Let us show . Similarly, the definitions of $x_{S_n}$ , $r_n$ , $I_n$ , $P_n$ and $\delta_n$ imply $$\begin{aligned} x_{S_n}+r_n&=\frac{\delta_n+y_n}{2}\cdot\frac{I_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &\text{\;\;\;\;}+\frac{1}{2}\cdot\frac{\delta_n P_n-y_n I_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\frac{\delta_n(I_n+P_n)}{2(1+\tan(A_{n+1}+B_{n+1}))}\\ &=\frac{2\delta_n(1+\tan(A_{n+1}+B_{n+1}))}{2(1+\tan(A_{n+1}+B_{n+1}))}\\ &=\delta_n.\end{aligned}$$ Now, we check . Likewise, the definitions of $y_{S_n}$ , $r_n$ , $I_n$ , $J_n$ , $P_n$ and $c_n$ yield $$\begin{aligned} y_{S_n}-r_n&=\frac{\delta_n+y_n}{2}\cdot\frac{J_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &\text{\;\;\;\;}-\frac{1}{2}\cdot\frac{\delta_n P_n-y_n I_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\frac{\delta_n(J_n-P_n)+y_n(I_n+J_n)}{2(1+\tan(A_{n+1}+B_{n+1}))}\\ &=\delta_n\tan(A_n+B_{n+1})+y_n\tan(A_n+B_{n+1})+y_n\\ &=(\delta_n+y_n)\cdot\tan(A_n+B_{n+1})+y_n\\ &=c_n.\end{aligned}$$ Lastly, we check . The definitions of $y_{S_n}$ , $r_n$ , $I_n$ , $J_n$ and $P_n$ imply $$\begin{aligned} y_{S_n}+r_n&=\frac{\delta_n+y_n}{2}\cdot\frac{J_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &\text{\;\;\;\;}+\frac{1}{2}\cdot\frac{\delta_n\cdot P_n-y_n\cdot I_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\frac{\delta_n\tan(A_{n+1}+B_{n+1})(1+\tan(A_n+B_{n+1}))}{1+\tan(A_{n+1}+B_{n+1})}\\ &+\frac{y_n(\tan(A_n+B_{n+1})\tan(A_{n+1}+B_{n+1})-1)}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\frac{\delta_n\tan(A_{n+1}+B_{n+1})(1+\tan(A_n+B_{n+1}))}{1+\tan(A_{n+1}+B_{n+1})}\\ &+\frac{y_n((1+\tan(A_n+B_{n+1}))(1+\tan(A_{n+1}+B_{n+1}))-I_n)}{1+\tan(A_{n+1}+B_{n+1})}\\ &=\frac{(\delta_n+y_n)(1+\tan(A_n+B_{n+1}))(1+\tan(A_{n+1}+B_{n+1}))}{1+\tan(A_{n+1}+B_{n+1})}\\ &+\frac{-\delta_n(1+\tan(A_n+B_{n+1}))-y_nI_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=(\delta_n+y_n)(1+\tan(A_n+B_{n+1}))\\ &+\frac{-\delta_n(1+\tan(A_n+B_{n+1}))-y_nI_n}{1+\tan(A_{n+1}+B_{n+1})}\\ &=c_n+(\delta_n-\delta_n'),\end{aligned}$$ where in the last equality we use the definitions of $c_n$ and $\delta_n'$ . 2. Recall, $$\begin{aligned} \Delta_{n}^{\varepsilon}=\{(y_1,y_2): &y_1=\varepsilon, (\varepsilon+1)\cdot\tan(A_n+B_n)+1<\\ &<y_2<(\varepsilon+x_n)\cdot\tan(A_n+B_{n+1})-1\}.\end{aligned}$$ We start by proving the inequality $$\begin{aligned} (\varepsilon+1)\cdot\tan(A_n+B_n)+1\leq c_n,\end{aligned}$$ for some $\delta_n'<\varepsilon<\delta_n$. Let $\delta_n'<\varepsilon<\delta_n$. Then the inequality $$\begin{aligned} (\varepsilon+1)\cdot\tan(A_n+B_n)+1\leq (\delta_n+1) \tan(A_n+B_n)+1\end{aligned}$$ holds. Therefore, the definition of $P_n'$ yields $$\begin{aligned} &(\varepsilon+1)\cdot\tan(A_n+B_n)+1\\ &\leq \delta_n\cdot\tan(A_n+B_{n+1})+(1+\tan(A_n+B_n)-\delta_n P_n').\end{aligned}$$ The definition of $y_n$ implies $$\begin{aligned} (\varepsilon+1)\cdot\tan(A_n+B_n)+1\leq (\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n.\end{aligned}$$ Eventually, the definition of $c_n$ implies $$\begin{aligned} (\varepsilon+1)\cdot\tan(A_n+B_n)+1\leq c_n,\end{aligned}$$ and hence the inequality. In order to prove the inequality $$\begin{aligned} c_n+(\delta_n-\delta_n')\leq(\varepsilon+x_n)\cdot\tan(A_n+B_{n+1})-1,\end{aligned}$$ it is enough to show it for $\varepsilon=\delta_n'$. We also use two relations: $$\begin{aligned} &\delta_n-\delta_n'=\delta_n'\cdot\tan(A_{n+1}+B_{n+1})-\delta_n\cdot\tan(A_n+B_{n+1})-y_n\cdot I_n,\label{del}\\ &y_n=\frac{\delta_n P_n-x_n\tan(A_n+B_{n+1})+1}{1+\tan(A_{n+1}+B_{n+1})}.\label{ips}$$ Notice that these statements follow directly from the definition of $\delta_n'$ and the equality $$\begin{aligned} d_n=(\delta_n+x_n)\cdot\tan(A_n+B_{n+1})-1\end{aligned}$$ from the Lemma \[lema4\], respectively. Due to $$\begin{aligned} &\delta_n'<\delta_n,\end{aligned}$$ and the definition of $P_n$ , the inequality $$\begin{aligned} &-\delta_n P_n + x_n\cdot\tan(A_n+B_{n+1})-1+\delta_n'\cdot\tan(A_{n+1}+B_{n+1})\\ & \leq(\delta_n'+x_n)\cdot\tan(A_n+B_{n+1})-1\end{aligned}$$ holds. Therefore, the equality implies $$\begin{aligned} &\delta_n'\cdot\tan(A_{n+1}+B_{n+1})-y_n\cdot(1+\tan(A_{n+1}+B_{n+1}))\\ &\leq (\delta_n'+x_n)\cdot\tan(A_n+B_{n+1})-1.\end{aligned}$$ From here, using , the inequality $$\begin{aligned} &(\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n+(\delta_n-\delta_n')\\ &\leq(\delta_n'+x_n)\cdot\tan(A_n+B_{n+1})-1\end{aligned}$$ follows. The definition of $c_n$ yields, $$\begin{aligned} &(\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n+(\delta_n-\delta_n')\\ &\leq(\delta_n'+x_n)\cdot\tan(A_n+B_{n+1})-1,\end{aligned}$$ and the inequality is proved. \[har1\] Let $I_n$, $P_n$, $P_n'$, $\delta_n$, $x_n$, $y_n$, $c_n$, $d_n$, $U_n$, $\delta_n'$, $S_n=(x_{S_n}, y_{S_n})$, $r_n$, $y_{P_1}$ and $ y_{P_2}$ be as in the Proposition \[prop5\]. Then 1. $y_{P_1}\leq c_n$, 2. $c_n + (\delta_n-\delta_n')\leq y_{P_2}$. <!-- --> 1. The inequality $$\begin{aligned} \varepsilon<\delta_n,\end{aligned}$$ implies $$\begin{aligned} (\varepsilon + y_n)\cdot \tan(A_n+B_{n+1})+y_n \leq (\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n.\end{aligned}$$ Therefore, together with and , the inequality $$\begin{aligned} y_{P_1}\leq c_n\end{aligned}$$ follows. 2. To prove the second inequality, we use $\delta_n'<\varepsilon$ and . Namely, $$\begin{aligned} \delta_n'\leq \varepsilon\end{aligned}$$ implies $$\begin{aligned} &\delta_n'\tan(A_{n+1}+B_{n+1}))-y_n(1+\tan(A_{n+1}+B_{n+1}))\\ &\leq \varepsilon\cdot\tan(A_{n+1}+B_{n+1})-y_n(1+\tan(A_{n+1}+B_{n+1})).\end{aligned}$$ From here we obtain $$\begin{aligned} &(\delta_n + y_n)\cdot\tan(A_n+B_{n+1})+y_n +\\ &+\delta_n'\cdot\tan(A_{n+1}+B_{n+1})-\delta_n\cdot \tan(A_n+B_{n+1})-y_n\cdot I_n\leq\\ &\leq (\varepsilon-y_n)\cdot\tan(A_{n+1}+B_{n+1})-y_n\end{aligned}$$ Using , and it follows $$\begin{aligned} c_n + (\delta_n-\delta_n')\leq y_{P_2},\end{aligned}$$ and hence the Lemma. \[har\] Let $I_n$, $P_n$, $P_n'$, $\delta_n$, $y_n$, $b_n$ and $M$ be as in the proof of Theorem \[har\_thm\_mat\], i.e. from Proposition \[prop4\]. Then: $$\begin{aligned} \limsup_{n\rightarrow \infty} \frac{M}{y_n}<\infty.\end{aligned}$$ By the definition of $\delta_n$ , it follows: $$\begin{aligned} y_n&=\frac{1+\tan(A_n+B_n)-\delta_n\cdot P_n'}{1+\tan(A_n+B_{n+1})}\notag\\ &=\frac{P_n^2\cdot(1+\tan(A_n+B_n))}{H_n},\label{y_n_g}\end{aligned}$$ where $H_n$ is like in the proof of the Proposition \[prop4\]. From here and it follows $$\begin{aligned} \label{komb1} &\frac{\delta_n}{y_n}=\frac{ I_n\cdot(1+\tan(A_{n+1}+B_{n+1}))}{P_n^2}.\end{aligned}$$ Also, by and the definitions of $\delta_n$ and $x_n$ , it follows $$\begin{aligned} \label{komb2} \frac{x_n}{y_n}=R_1+R_2+R_3,\end{aligned}$$ where $$\begin{aligned} R_1&=2\cdot\frac{1+\tan(A_n+B_{n+1})}{\tan(A_n+B_{n+1})}\cdot\frac{1+\tan(A_{n+1}+B_{n+1})}{P_n},\\ R_2&=\frac{1+\tan(A_n+B_{n+1})}{(1+\tan(A_n+B_n))\cdot\tan(A_n+B_{n+1})},\\ R_3&=\frac{I_n\cdot(1+\tan(A_{n+1}+B_{n+1}))}{P_n^2} \frac{P_n'}{\tan(A_n+B_{n+1})}\frac{1}{1+\tan(A_n+B_n)}.\end{aligned}$$ Combining (\[komb1\]) and (\[komb2\]), the conditions (\[uvjet\]) and (\[uvjet\_d\]) imply $$\begin{aligned} \limsup_{n\rightarrow \infty} \frac{M}{y_n}<\infty,\end{aligned}$$ and hence the proposition. [10]{} K. Bogdan, *The boundary Harnack principle for the fractional Laplacian*, Studia Math. 123 (1997), no.1, 43–80. MR 1438304 (98g:31005) K. Bogdan and P. Sztonyk, *Harnack’s inequality for stable [L]{}évy processes*, Potential Anal. 22 (2005), no.2, 133–150. MR 2137058 (2005m:31015) K. Bogdan and P. Sztonyk, *Estimates of the potential kernel and [H]{}arnack’s inequality for the anisotropic fractional [L]{}aplacian*, Studia Math. 181 (2007), no.2, 101–123. MR 2320691 (2008k:31015) Z.-Q. Chen, *Lecture Notes: Potential Theory for Jump Processes; Summer School 2012 on Nonlocal Operators*, 2012. N. Ikeda and S. Watanabe, *On some relations between the harmonic measure and the [L]{}évy measure for a certain class of [M]{}arkov processes*, J. Math. Kyoto Univ. 2 (1962), 79–95. MR 0142153 (25 \#5546) T. Jakubowski, *On Harnack inequality for $\alpha$-stable Ornstein-Uhlenbeck processes*, Math. Z. 258 (2008), no. 3, 609–628. MR 2369047 (2009b:60242) N. S. Landkof, *Foundations of modern potential theory*, Die Grundlehren der mathematischen Wissenschaften in Einzeldarstellungen mit besonderer Berücksichtigung der Anwendungsgebiete ; 180, vol. 180, Springer, Berlin \[u.a.\], 1972. E. Priola, *Pathwise uniqueness for singular [SDE]{}s driven by stable processes*, Osaka J. Math. 49 (2012), no. 2, 421–447. MR 2945756 K. Sato, *Lévy processes and infinitely divisible distributions*, digital print. ed., Cambridge studies in advanced mathematics; 68, vol. 68, Cambridge Univ. Press, Cambridge \[u.a.\], 2007 (eng). P. Sztonyk, *Estimates of tempered stable densities*, J. Theoret. Probab. 23 (2010), no. 1, 127–147. MR 2591907 (2011b:60183) S. J. Taylor, *Sample path properties of a transient stable process*, J. Math. Mech. 16 (1967), 1229–1246. MR 0208684 (34 \#8493) [^1]: The author appreciates the support of the International Graduate College “Stochastics and Real World Models”, Universität Bielefeld. -
- Assistant
To process the given complex mathematical text from the article "HARNACK INEQUALITIES FOR SYMMETRIC STABLE LÉVY PROCESSES," we need to differentiate between parts that require quick, straightforward responses (System 1 thinking) and parts requiring in-depth, reflective thinking (System 2 thinking). Here's how we can structure it:
[sys1]
The article discusses Harnack inequalities in relation to symmetric -stable Lévy processes in .
The focus is on whether these inequalities hold under specific conditions related to the Lévy measure.
[end sys1][sys2]
2.1 Understand the Problem:
The main inquiry is whether the Harnack inequality holds for a symmetric -stable Lévy process under certain conditions. This involves examining the behavior of harmonic functions within specified sets.Reflection:
"What are the key constraints limiting the applicability of the Harnack inequality? How do the specific characteristics of the -stable Lévy process influence this analysis?"Creative Perspective:
"Could there be parallel mathematical frameworks where similar inequalities hold or break down, providing a comparative insight?"2.2 Analyze the Information:
The crucial elements include the measure , the construction of harmonic functions, and the spectral properties influencing inequality outcomes.Reflection:
"Are there assumptions made about the distribution or symmetry that could skew interpretations? Could external factors like dimensionality cause deviations?"Creative Perspective:
"Exploring how minor variations in the measure could lead to different conclusions about the inequality.”2.3 Generate Hypotheses:
- The inequality holds for approaching 2 but fails as nears 0. (Confidence: 0.7, Creative: 0.6)
- Non-degenerate measures are critical for the inequality to hold. (Confidence: 0.8, Creative: 0.5)
- The relation between spectral measure and surface uniformity impacts inequality adherence. (Confidence: 0.6, Creative: 0.7)
- Alternative probability kernels might reveal conditions under which inequalities hold. (Confidence: 0.5, Creative: 0.8)
- The influence of dimensionality significantly shifts eligibility of inequalities. (Confidence: 0.6, Creative: 0.6)
- Finite limit behavior of processes enhances inequality validity. (Confidence: 0.7, Creative: 0.7)
- External interactions, like environmental changes, affect process trajectories and subsequently inequalities. (Confidence: 0.4, Creative: 0.9)
- Specific boundary conditions strongly dictate outcomes related to inequalities. (Confidence: 0.7, Creative: 0.5)
- Comparative studies with other stochastic processes might offer fresh insights into inequality dynamics. (Confidence: 0.5, Creative: 0.9)
- Varying the density function could systematically test inequality boundaries. (Confidence: 0.6, Creative: 0.8)
Reflection:
"Have we accounted for all known interactions? What unconventional paths could help refine these hypotheses?"Creative Perspective:
"Leveraging computational simulations to identify unseen stability markers or anomalies."2.4 Anticipate Future Steps and Obstacles:
Exploration into related inequalities requires simulations under varying conditions for generalization.Reflection:
"What specific challenges might skew the data? Can current methodologies adapt to dynamism in mathematical modeling?"Creative Perspective:
"Foresee new computational tools emerging that could simplify inequality examinations."[end sys2]
[sys1]
To summarize, investigations build on the Harnack inequality's interactions with the -stable Lévy processes, emphasizing boundary conditions and measure properties. Future focus should include simulations for confirming hypotheses and identifying systemic patterns or discrepancies.
[end sys1]