A note on maximal operator on \(\ell^{p_n}\) and \(L^{p(x)}(\mathbb R)\) (Q2471299): Difference between revisions

As shown by the author in [Math.\ Inequal.\ Appl.\ 7, No.\,2, 255--265 (2004; Zbl 1059.42016)], the discrete space \(\ell^{\{p_n\}}\) with variable exponent \(p_n\) is equivalent to the space \(\ell^p\) with constant \(p\) if and only if there exists a \(c>0\) such that \[ \sum_{n: |p_n-p|>0}|p_n-p|c^\frac{1}{|p_n-p|}<\infty; \] this condition is called the \(\mathcal{P}\) condition. The author shows that for the boundedness of the maximal operator in the space \(\ell^{p_n}\), condition \(\mathcal{P}\) is not necessary, but may not be omitted in general, by constructing two examples. In the first example, only the condition \(\lim_{n\to\infty}p_n=p\) is satisfied and the maximal operator is not bounded. In the second example, the condition \(\mathcal{P}\) is not fulfilled and the maximal operator is unbounded. Similarly, the author constructs an example of a function \(p(x)\) such that the known integral condition sufficient for the boundedness of the maximal operator in \(L^{p(\cdot)}(\mathbb{R}^n)\) (also due to the author) is not satisfied, nevertheless the maximal operator is bounded.