We consider linear difference equations whose coefficients are meromorphic at $\infty$. We characterize the meromorphic equivalence classes of such equations by means of a system of meromorphic invariants. Using an approach inspired by the work of G. D. Birkhoff we show that this system is free.

In the paper a modification of Samoilenko's numerical analytic method is adapted for solving of boundary value problems for difference equation. Similarly to the case of differential equations it is shown that the considered modification of the method requires essentially less restrictive condition-then the original method-for existence and uniqueness of solution of auxiliary equations which play a crucial role in solving the boundary value problems for difference equations.

We show that, generally, families of measurable functions do not have the difference property under some assumption. We also show that there are natural classes of functions which do not have the difference property in ZFC. This extends the result of Erdős concerning the family of Lebesgue measurable functions.

Let $a₁,...,a_k$ be arbitrary nonzero real numbers. An $(a₁,...,a_k)$-decomposition of a function f:ℝ → ℝ is a sum $f₁+\cdots +f_k=f$ where $f_i:ℝ\to ℝ$ is an $a_i$-periodic function. Such a decomposition is not unique because there are several solutions of the equation $h₁+\cdots +h_k=0$ with $h_i:ℝ\to ℝa_i$-periodic. We will give solutions of this equation with a certain simple structure (trivial solutions) and study whether there exist other solutions or not. If not, we say that the $(a₁,...,a_k)$-decomposition is essentially unique. We characterize those periods for which essential uniqueness...

We consider Schrödinger operators with dynamically defined potentials arising from continuous sampling along orbits of strictly ergodic transformations. The Gap Labeling Theorem states that the possible gaps in the spectrum can be canonically labelled by an at most countable set defined purely in terms of the dynamics. Which labels actually appear depends on the choice of the sampling function; the missing labels are said to correspond to collapsed gaps. Here we show that for any collapsed gap,...

Let $ℬ=f:ℝ\to ℝ:fisbounded$, and $ℳ=f:ℝ\to ℝ:fisLebesguemeasurable$. We show that there is a linear operator $\Phi :ℬ\to ℳ$ such that Φ(f)=f a.e. for every $f\in ℬ\cap ℳ$, and Φ commutes with all translations. On the other hand, if $\Phi :ℬ\to ℳ$ is a linear operator such that Φ(f)=f for every $f\in ℬ\cap ℳ$, then the group $G_{\Phi }$ = a ∈ ℝ:Φ commutes with the translation by a is of measure zero and, assuming Martin’s axiom, is of cardinality less than continuum. Let Φ be a linear operator from ${ℂ}^{ℝ}$ into the space of complex-valued measurable functions. We show that if Φ(f) is non-zero for every $f\left(x\right)=e^{cx}$, then $G_{\Phi }$ must...