Halasz G., Lovasz L., Simonovits M., Sos V.T. (редакторы). Пол Эрдос и его математика, издание 1 (Springer, 2002), Halasz G., Lovasz L., Simonovits M., Sos V.T. (eds.) Paul Erdos and his mathematics, vol.1 (Springer, 2002)(KA)(600dpi)(T)(734s).djvu:

Halasz G., Lovasz L., Simonovits M., Sos V.T. (редакторы). Пол Эрдос и его математика, издание 1 (Springer, 2002)

Halasz G., Lovasz L., Simonovits M., Sos V.T. (eds.) Paul Erdos and his mathematics, vol.1 (Springer, 2002)(KA)(600dpi)(T)(734s).djvu

Size 7.8Mb
Date Jan 27, 2006

Cites: At
the same time Erdos proved that the analogous statement for measurable
functions is false, at least if we suppose the continuum hypothesis (we shall
discuss Erdos' example in Section 6)...
As we saw above, we have (LS) => (DDP) and (DP) => (DDP)
for every Abelian topological group and Banach space B...
It is easy to check, by a small modification of the proof, that the
following version of Theorem 2.1 is true...
Let p be a polynomial such that p(i) = f(i) for i = 0,1,2; p(l/2) >
/A/2); and pC/2) < /C/2)...
A linear map M : Loo(G,J5) -» В is called a B-valued
mean, if (i) \M(f)\ < l/l^ for every / E L^{G,B), and (ii) / = b E В
implies M(f) = b...
Since
D is dense in Г, it is easy to see that whenever p is an extension of q \ E
to R then ш(р | F,x) > u(q \ E,x) for every x E F...
Therefore G is
Lipschitz in (a-7/, a+7/), and thus G{x)-G{a) = /ax Gf{t) dt in (a-7/, a4-r/)...
Zahorski proves that Mo = M\ D M2 D Мг D M4 D M$, the class
.Д^о = M\ equals the class of Darboux Baire 1 functions, and M$ equals the
class of approximately continuous functions...
Let n > 0, and suppose that we have defined xo,..., xn € U and yo,..., yn €
У such that A0) holds for every 0 < i < j < n, and A1) holds for every
1 < i < n...
Since
g(n(x - y)) - 5@) - n(g(x) - g(y))
г=1
we have | g(n(x - y)) - 5@) - n(g(x) - g{y)) | < n0(x - y) and
n
for every n...
If e > 0, let J6jn
be the number of intervals of the form
with
The routine estimates for one dimensional Brownian motion (details omit-
omitted) can be used to show that for every e > 0
E[Je>n] < Ce,
E[Je>n|J2e>n>0]>Celogn,
414 G, F...
While we still do not know the value of
this exponent, the rigorous bounds to tell us that the dimension is strictly
1 Note added in proof: This has recently been proved by Lawler, Schramm and
Werner...
The fainthearted would have concluded from Runge and Faber-Bern-
stein's work that interpolation by polynomials is an inherently flawed pro-
process...
Almost every course in func-
functional analysis contains a treatment of orthonormal expansions in Hilbert
space and the least squares property of their partial sums...
(This again reminds one that the importance of a sub-
subject is often measured by its external impact...
They also strengthened the convergence A4) in terms of errors of best
approximation, defined above...
An immediate
question is to what extent the convergence in Lp persists for all, or some,
p...
We apply Lemma 2 with Л = {p + 2 : p < x}, z = я\ f2 =
xl/2(logx)-c, X = liz, V = P\ {2}, u{p) = ^j...
The main efforts
are paid to survey the results on the weak convergence of processes defined in
terms of arithmetical functions...
Actually,
linear combinations of additive functions were investigated in [41] but we
leave relevant problems for the next sections...
Our proof uses some elementary properties of the prime numbers and
their distribution and some classical results from diophantine approxima-
approximation...
Thus, the question naturally
arises as to how effective a construction of an n-point set can be, see [40],
p...
Also, the corresponding statement for R2, using sets of lines L\ and L2 and
sets Si and S2 is false...
U Sp such that every line in Li meets
Si in a set of size < u)e-p
I note that the fact that A.3) implies A.1) was proven earlier by Roy
Davies [8] and that some of the key ideas of our arguments go back to
combinatorial arguments of Erd6s and Hajnal [13]...
Hence one might
expect that numerical studies of primes in short intervals would lend support
to the conjectural relation A)...