Some properties and interpolation theorems in weak Orlicz-Lorentz spaces (Q2043674): Difference between revisions

Define \[\Vert f\Vert_{\Lambda_X^{\varphi,\infty}(w)}=\inf\Bigl\{c>0:\, \sup_{t>0}\varphi\Bigl(\frac{f^\ast(t)}{c}\Bigr)W(t)\leq 1 \Bigr\}\] where \(w\geq 0\), \(W(t)=\int_0^t w\), and \(f^\ast\) is the decreasing rearrangement of \(f\) on the \(\sigma\)-finite measure space \((X,\mu)\). Here, \(\varphi\) is an increasing, convex function on \((0,\infty)\) such that \(\varphi(0)=0\), \(\varphi(2t)\leq C\varphi(t)\), and \(\varphi(\infty)=\infty\). The weak Orlicz-Lorentz space \(\Lambda_X^{\varphi,\infty}(w)\) is the set of measurable \(f\) such that \(\Vert f\Vert_{\Lambda^{\varphi,\infty}(w)}<\infty\) (reference to \(X\) is suppressed).\par Several convergence and embedding properties of the spaces \(\Lambda_X^{\varphi,\infty}(w)\) are established. For example, if \(w\in L^1\) and \(0<q<q_\varphi<p_\varphi<p\leq \infty\) then \(\Lambda^p(w)\subset \Lambda^{\varphi,\infty}(w)\subset \Lambda^q(w)\), where \(\Lambda^p(w)\) is the classical Lorentz space defined by \(\Vert f\Vert_{\Lambda_X^p(w)}=\Vert f^\ast\Vert_{L^p(w)}<\infty\), and \(q_\varphi=\inf_{t>0}\frac{t\varphi(t)}{\varphi(t)}\). \par It is proved (Thm.~2.6) that if \(w\in L^1\) and \(1<q_\varphi<\infty\) then \(\Lambda^{\varphi,\infty}(w)\) is normable. It is also proved that if \(\Vert f-f_n\Vert_{\Lambda^{\varphi,\infty}(w)}\to 0\) then \(f_n\to f\) in measure (Thm.~3.3)). One has the Fatou-type bound \(\Vert \lim\inf_{n\to\infty}\vert f_n\vert \Vert_{\Lambda^{\varphi,\infty}(w)}\leq \lim\inf_{n\to\infty}\Vert f_n\Vert_{\Lambda^{\varphi,\infty}(w)}\) (Thm.~3.5). Proposition 3.7 establishes several criteria equivalent to \(\Lambda^{\varphi,\infty}(w)\) having an absolutely continuous norm, meaning that \(\lim_{n\to\infty} \Vert f\chi_{A_n}\Vert_{\Lambda^{\varphi,\infty}(w)}=0\) whenever \(\{A_n\}\) is a sequence of measurable sets such that \(\chi_{A_n}\) decreases to zero almost everywhere.\par Finally, some interpolation properties are established. Let \(0<p_0<q_\varphi\leq p_\varphi<p_1<\infty\). Let \(w\) and \(w_1\) be two decreasing weights and let \(T\) be a quasi-linear operator defined on \(\Lambda^{p_0}(w)+\Lambda^{p_1}(w_1)\). Suppose that \(T:\lambda^{p_i}(w)\to \Lambda^{p_i,\infty}(w_1)\) is bounded (\(i=0,1\)). Then \(T:\Lambda^{\varphi,\infty}(w)\to \Lambda^{\varphi,\infty}(w_1)\) is bounded (Thm.~4.5). If, in addition, \(T:\Lambda^{p_0}(w)\to \Lambda^{p_0,\infty}(w_1)\) is bounded and \(T:L^\infty\to L^\infty\) is bounded, then \(T:\Lambda^{\varphi,\infty}(w)\to \Lambda^{\varphi,\infty}(w_1)\) is bounded (Thm.~4.6.). A third result establishes boundedness of \(T:\Lambda^{\varphi}(w)\to \Lambda^{\varphi}(w_1)\) assuming boundedness from \(\Lambda^{p_i}(w)\) to \(\Lambda^{p_i,\infty}(w_1)\), \((i=0,1\)).