On multiplicative functions which are small on average

Let f be a completely multiplicative function that assumes values inside the unit disc. We show that if ${\sum_{n \leq x}f(n)\ll x/(\rm log x)^A}$ ∑ n ≤ x f ( n ) ? x / ( l o g x ) A , ${x \geq 2}$ x ≥ 2 , for some A > 2, then either f(p) is small on average or f pretends to be ${\mu(n)n^{it}}$ μ ( n ) n i t for some t.     

On multiplicative functions which are additive on sums of primes

Let ${f : \mathbb{N} \to \mathbb{C}}$ be a multiplicative function satisfying f(p ₀) ≠ 0 for at least one prime number p ₀, and let k ≥ 2 be an integer. We show that if the function f satisfies f(p ₁ + p ₂ + . . . + p _k ) = f(p ₁) + f(p ₂) + . . . + f(p _k ) for any prime numbers p ₁, p ₂, . . . ,p _k then f must be the identity f(n) = n for each ${n \in \mathbb{N}}$ . This result for k = 2 was established earlier by Spiro, whereas the case k = 3 was recently proved by Fang. In the proof of this result for k ≥ 6 we use a recent result of Tao asserting that every odd number greater than 1 is the sum of at most five primes.     

On improving distribution function estimators which are not monotonic functions

In this paper we consider the problem of correcting distribution function estimators which are not nondecreasing functions (for example kernel type estimators). The method is based on the orthogonal projection in L₂ and guarantees improving of the integrated mean square error for each sample size.     

On the existence of Pettis integrable functions which are not Birkhoff integrable

Let be a weakly Lindelöf determined Banach space. We prove that if is non-separable, then there exist a complete probability space and a bounded Pettis integrable function that is not Birkhoff integrable; when the density character of is greater than or equal to the continuum, then is defined on with the Lebesgue measure. Moreover, in the particular case (the cardinality of being greater than or equal to the continuum) the function can be taken as the pointwise limit of a uniformly bounded sequence of Birkhoff integrable functions, showing that the analogue of Lebesgue's dominated convergence theorem for the Birkhoff integral does not hold in general.

     

On optimizing functions which are defined in part by an approximation process

This paper studies an optimization problem in which the objective function can not be completely given in closed form. In particular, we assume that some part of the objective function must be computed by an approximation process. This paper develops a technique for solving a class of such problems. Examples demonstrating the technique and problem areas in which it has been successfully applied are also given.     

On the univalence of derivatives of functions which are univalent in angular domains

We consider functions f that are univalent in a plane angular domain of angle απ, 0 < α ≤ 2. It is proved that there exists a natural number k depending only on α such that the kth derivatives f ^(k) of these functions cannot be univalent in this angle. We find the least of the possible values of for k. As a consequence, we obtain an answer to the question posed by Kir’yatskii: if f is univalent in the half-plane, then its fourth derivative cannot be univalent in this half-plane.     

On decoupling of functions of normal vectors

Two decoupling type inequalities for functions of Gaussian vectors are proved. In both cases, it turns out that the case of linear functions is the extreme one. The proofs involve certain properties ofWick’s (Hermite’s) polynomials and a refined version of Schur’s theorem on entrywise product of positive definite matrices.     

Systems of functions which are complete in measure

In this note it is proved that if a complete orthonormal system {? _n} in L₂[0, 1] contains a subsystem {? _nk} of a lacunary order p>2, then for some bounded measurable function h(x) the system {h(x)? _n(x)}_n≠_nk is complete in L₂[0, 1].     

On typically real functions which are generated by a fixed typically real function

Let T be the family of all typically real functions, i.e. functions that are analytic in the unit disk Δ:= {z ∈ ℂ: |z| < 1}, normalized by f(0) = f′(0) − 1 = 0 and such that Imz Im f(z) ⩾ 0 for z ∈ Δ.     

Approximately continuous functions which are continuous almost everywhere

Acta Mathematica Hungarica -     

On sums which are powers

The cardinality of sets A is estimated under the conditions that every element of the sum set A+A is a power resp. powerful number (n is said to be powerful if p n implies p ² n). Subset sums with these properties are also studied. This revised version was published online in June 2006 with corrections to the Cover Date.     

On optimizing certain nonlinear convex functions which are partially defined by a simulation process

This paper considers the problem of optimizing a nonlinear convex function which is, in part, defined by a simulation process. An iterative technique, based on contractive mapping properties, is given by which a selected class of such functions can be optimized.