Searcharxiv⌕ Search

arXiv subjects

Hailu Bikila Yadeta

Publications and source records attributed to Hailu Bikila Yadeta.

10 recordsLinked to original sources

Decomposition of Spaces of Periodic Functions into Subspaces of Periodic and Antiperiodic Functions and Its Connection to the Rademacher System and the Haar Wavelet Basis

We prove that the space $\mathbb{P}_p$ of $p$-periodic functions decomposes as the direct sum $\mathbb{P}_{p/2} \oplus \mathbb{A}\mathbb{P}_{p/2}$, where $\mathbb{P}_{p/2}$ denotes the space of functions periodic with period $p/2$ and $\mathbb{A}\mathbb{P}_{p/2}$ denotes the space of functions antiperiodic with antiperiod $p/2$ (i.e., $f(x+p/2) = -f(x)$). Iterating this decomposition yields a hierarchy of refined periodic subspaces. Under suitable uniform decay conditions on the residual periodic components, any $p$-periodic function on a compact interval admits a convergent expansion into a series of antiperiodic components with distinct antiperiods. As a concrete example, the continued periodic-antiperiodic decomposition of the fractional part function $\{x\}$ generates the Rademacher system. Additionally, we examine an orthogonal decomposition of $L^2(0,1)$ induced by reflection symmetry about the midpoint $x = 1/2$, i.e., $f(x) = \pm f(1-x)$. Using explicit projection operators, we show that this reflection-based decomposition generates a multiscale structure analogous to the Haar multiresolution analysis: the antiperiodic (odd-reflection) component yields a system equivalent to the Haar wavelet family $\{ψ_{j,k}\}$, while the periodic (even-reflection) component corresponds to the scaling space of piecewise constant functions. This provides a boundary-condition-based interpretation of the Haar wavelet basis.

math.GM↗

Initial Value Problems for Some Functional Equations and the PENLP

In this paper, we study initial value problems for a class of functional equations. We introduce the concept of appropriate initial sets to enable the unique extension of an initial function into a solution defined over larger domains. Our analysis characterizes the structure and size of these initial sets, illustrating how functions specified on them can be uniquely extended to broader regions. Furthermore, we propose the Principle of Exclusion of Neighborhoods of Limit Points (PENLP)-a novel conceptual framework that elucidates the behavior of solutions near limit points and their impact on the extension process. We demonstrate that the initial sets must avoid the limit points of the functional equations to ensure uniqueness and well-posedness. The functional equations examined in this study include: $$ y(x+1)= y(bx),\quad b \neq 0, $$ $$ y(x)=y(2x),\quad y(x)=y(-2x),\quad y(3x)=y(x)+y(2x), $$ as well as the equations characterizing even functions and odd functions.

math.GM↗

Analysis of the Distribution and Asymptotic Approximations of Roots of the Polynomial Equation $$ z^{n+1}=(1+z)^n, n \in \mathbb{N} $$ in the Complex Plane

We study the spatial distribution of the positive, negative and non-real complex roots $z_n $ of the sequence the $(n+1)$th degree polynomial equation $$ z^{n+1}=(1+z)^n,\quad n \in \mathbb{N}.$$ We establish asymptotic approximations to the sequence of the negative, the positive and the non-real complex roots of the equation as $n\rightarrow \infty $. In addition, we discuses the possible areas of applications of the current problem.

math.CV↗

Equality of the Casoratian and the Wronsikian for some Sets of Functions and an Application in Difference Equations

The Casoratian determinants are very important in the study of linear difference equation, just as the Wronskian determinants are very important in the study of linear ordinary differential equations. The Casoratian and Wronskian determinants of a given set of functions are generally different. In this paper, we show that the Wroskian $ W[1, x, x^2,...,x^n ]$ and the Casoratian $ C[1,x, x^2,...,x^n ] $ of the set of functions $ \{1, x, x^2,...,x^n: \, n \in \mathbb{N} \}$ are equal to each other and independent of the variable $x$. Furthermore, we show that $ W[p_1, p_2,...,p_n, p_{n+1} ]= C[p_1, p_2,..., p_n, p_{n+1} ]$ for any basis $ \mathcal{B} = \{ p_1, p_2,..., p_n, p_{n+1} \}$ of the vector space spanned by $ \{ 1, x, x^2,...,x^n \}$. We also explore applications of such Casoratians in deriving the general solutions of certain classes of homogeneous linear difference equations.Additionally, the paper discusses scenarios where the Casoratian and Wronskian are proportional, as well as cases where they differ for finite sets of functions.

math.GM↗

On the Generalization of Weinberger's Inequality with Alternating Signs

For given set of $m$ positive numbers satisfying the conditions: $$ a_1 \geq a_2 \geq , ... \geq a_m \geq 0, $$ the inequality $$ \sum_{s=1}^{m} (-1)^{s-1}a^r_s \geq \left[ \sum_{s=1}^{m} (-1)^{s-1}a_s\right]^r, \quad r > 1, $$ was proved by H. Weinberger. The generalization of Weinberger's result takes the form $$ \sum_{s=1}^{m} (-1)^{s-1}f(a_s) \geq f\left( \sum_{s=1}^{m} (-1)^{s-1}a_s\right), $$ where $f$ is a convex function satisfying the condition $f(0)\leq 0 $. The condition $f(0)\geq 0 $ in the generalization proposed by Bellman was corrected by Olkin as $f(0) \leq 0 $. Bellman gave only a graphical proof for differentiable convex functions. In this paper, we give a mathematical proof for the generalized inequality including the importance of the condition $f(0)\leq 0$. We introduce a set $\mathcal{W}$ of functions so that functions in the intersection of $\mathcal{W}$ and the set of all convex functions are the ones that are desirable in the generalization. In addition, we give a proof of Szegö's inequality which applies to sums with odd number of terms.

math.CA↗

New Indefinite Summation Formulas and Some Applications

In this paper, we introduce a novel indefinite summation $\sum_{t} f(t)$ (or antidifference $Δ^{-1}f(t) $ ) formula for any given function $f$. We apply the indefinite summation formula to calculate a particular solution to a nonhomogeneous linear difference equation of the form $$ y(x+h)-λy(x)=f(x),\quad h > 0,\quad λ\neq 0, $$ and also to solve a linear difference inequality of the form $$ y(x+h)-λy(x) \geq 0,\quad h > 0,\quad λ\neq 0. $$ Furthermore, we apply the formula to determine a particular solution to a difference equations of the form $$ Φ(E)y(t)=f(t), $$ and in solving a linear difference inequality of the form, $$ Φ(E)y(t)\geq 0, $$ where $ Φ(E) $ is some linear difference operator. We show how the antidifference of a function $f$ calculated with the current formula is related to the already existing result and establish the corresponding identity.

math.CA↗

Direct sum Decomposition of Spaces of Periodic Functions: $$ \mathbb{P}_n = \bigoplus \limits_{d|n} \ker(Φ_d(E))

It was proved that the space $ \mathbb{P}_p $ of all periodic function of fundamental period $ p $ is a direct sum of the space $ \mathbb{P}_{p/2} $ of all periodic functions of fundamental period $ p/2 $ and the space $ \mathbb{AP}_{p/2} $ of all antiperiodic functions of fundamental antiperiod $ p/2 $. In this paper, we study some connections between periodic functions, cyclotomic polynomials, roots of unity, circulant matrices, and some classes of difference equations. In particular, we state and prove the sufficient condition for the existence of periodic solutions of integer period or arbitrary period of some difference equation. We also show that the space $ \mathbb{P}_n $ of all periodic functions of integer period $n$ can be decomposed as the direct sum of operators' kernels $\ker (Φ_d(E)) $, where $Φ_d(E),\, 1 \leq d \leq n, d|n $ are the cyclotomic polynomials of the shift operator $E$. We state and prove important theorems, state and prove the necessary and sufficient conditions for a linear difference equation with constant coefficients to have periodic solutions.

math.NT↗

Analysis of the Differential-Difference Equation $y(x+1/2)-y(x-1/2) = y'(x)$

In this paper we study some solution techniques of differential-difference equation $$ y'(x) = y(x + 1/2)- y(x- 1/2),$$ first without an initial condition and then with some initial function $h$ defined on the unit interval $ [-1/2, 1/2]$. We show some sufficient conditions that an initial function $h$ is admissible, i.e., it yields a unique continuous solution on some symmetric interval about $0$.

math.CA↗

New characterizations of the ring of the split-complex numbers and the field $ \mathbb{C} $ of complex numbers and their comparative analyses

In this paper, we give a new characterization of the split-complex numbers as a vector space $LC_2= \{xI+yE : x,y \in \mathbb{R},\, E^2=I \}$ of operators, where $I$ is the identity operator and $E$ is the unit shift operator that are operating on the space $\mathbb{P}_2$ of all real-valued $2$-periodic functions. We also characterize the field of the complex number $\mathbb{C}= \{x+yi: x, y \in \mathbb{R}, i^2 =-1 \}$ as the space of linear operators of the form $ \{xI+yE, E^2 = -I \}$, where $I$ is the identity operator and $E$ the unit shift operator that are regarded as operating on the vector space $\mathbb{AP}_2 $ of all real-valued $2$-antiperiodic functions. In an analogy to the polar form of complex numbers, we form the hyperbolic form of some subset $ \mathcal{H} $ of the elements of $LC_2 $. We study some properties of the elements of $LC_2$, the trace, the determinant, invertibility conditions, and others. We study some elementary functions defined on subsets of $LC_2$ as compared and contrasted with the usual complex functions. We study properties like continuity, differentiability and define the holomorphic condition of $LC_2$ functions in a different sense than complex functions. We establish the line integrals of the vector-valued functions in $LC_2$ and compare them against the well known results for complex functions of a complex variable.

math.CV↗

Real Analytic Methods in the Formulations of some Combinatorial Inequalities

In this paper, we derive some new combinatorial inequalities by applying well known real analytic results like Hölder's inequality, Young's inequality, and Minkowiski's inequality to the recursively defined sequence $f_n$ of functions \begin{align*} f_0(x) & = χ_{(-1/2, 1/2)} (x), \nonumber f_{n+1}(x) & = f_n(x+1/2)+ f_n(x-1/2), n \in \mathbb{N}\,\cup \,\{0\}. \end{align*} Towards this goal, we derive the closed form of the aforementioned sequence $ (f_n)_{n\in \mathbb{N}\,\cup \,\{0\}}$ of functions and show that it is a sequence of simple functions that are linear combinations of characteristic functions of some unit intervals $ I_{n,i},\, i=0,1, ..., n $, with values the binomial coefficients $ \binom{n}{i}$ on each unit interval $I_{n,i}$. We show that $ f_n \in L^p(\mathbb{R})),\, 1\leq p \leq \infty $. Besides applying real analytic methods to formulate some combinatorial inequalities, we also illustrate the application of some combinatorial identities. For example, we use the Vandermonde convolution (or Vandermonde identity), in the study of some properties of the sequence of functions $ (f_n)_{n\in\mathbb{ N}\cup \{0\}}$. We show how the $L^2$ norm of $f_n$ is related to the Catalan numbers.

math.CO↗