Let $p\geq 1$ and let $(w_n)_{n\in\mathbb N}$ a non-increasing sequence of positive numbers such that: $$ \lim_{m\in\mathbb N} w_n = 0 \qquad \text{and}\qquad \sum_{n=1}^\infty w_n = \infty $$
We denote $S_\infty = \{ f:\mathbb N\to \mathbb N|\, f \text{ is a bijection}\}$. If $(x_n)_{n\in \mathbb N}$ is a sequence such that:
$$ \sup_{\pi\in S_\infty} \left( \sum_{n=1}^\infty |x_{\pi(n)}|^p w_n \right) <\infty, $$ we say that $(x_{\pi(n)}^*)_{n\geq 1}$ is a non-increasing rearrangement of $(x_n)_{n\geq 1}$ when there exists a permutation of integers, $\sigma$ such that $(x^*_n)_{n\geq 1}=(|x_{\sigma(n)}|)_{n\geq 1}$ is non-increasing.
I want to prove the following inequality:
Suppose that $(x_n)_{n\geq 1}$ is a sequence such that $\sup\limits_{\pi\in S_\infty} \left(\sum\limits_{n=1}^\infty |x_{\pi(n)}|^p\, w_n \right)^{\frac{1}{p}}\!\!<\infty$, then the following inequality holds true: $$ \sup\limits_{\pi\in S_\infty} \left(\sum\limits_{n=1}^\infty |x_{\pi(n)}|^p\, w_n \right)^{\frac{1}{p}}\!\!\leq \left(\sum\limits_{n=1}^\infty |x_n^*|^p\, w_n \right)^{\frac{1}{p}} $$
What I have done so far:
If $\sigma$ is a permutation such that $(x^*_n)_{n\geq 1}=(|x_{\sigma(n)}|)_{n\geq 1}$ and $\pi\in S_\infty$, if suffices to show that: $$ \sup\limits_{\pi\in S_\infty} \sum\limits_{n=1}^\infty |x_{\pi(n)}|^p\, w_n \leq \sum\limits_{n=1}^\infty |x_n^*|^p\, w_n $$
Each one of the series is the limit of an (increasing) sequence of positive terms that is bounded from above, therefore both limits are the suprema of the finite sums. If, for $k\in \mathbb N$ I can find $k'\in\mathbb N$ such that: $$ \sup\limits_{\pi\in S_\infty} \sum\limits_{n=1}^k |x_{\pi(n)}|^p\, w_n \leq \sum\limits_{n=1}^{k'} |x_n^*|^p\, w_n $$
then the desired inequality follows. The problem is that I can't find a good way to choose such $k'$. I tried to look for $\sigma \circ \pi^{-1}(j)$ for $1\leq j\leq k$ but I couldn't figure what to do with $(w_n)_{n\geq 1}$, since they are not affected by $\pi$ or $\sigma$.
I would appreciate if anyone can suggest an elegant argument to finish my proof or provide a different argument.