论文标题
混合对于无三角形的反身图很难
Mixing is hard for triangle-free reflexive graphs
论文作者
论文摘要
储层计算是预测湍流的有力工具，其简单的架构具有处理大型系统的计算效率。然而，其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间，然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用，该系统模拟了湍流的若干特征。我们表明，使用径向基函数作为非线性投影器，即使只有部分观测并且不知道控制方程，也能稳健地捕捉复杂的系统非线性。最后，我们表明，当测量稀疏、不完整且带有噪声，甚至控制方程变得不准确时，我们的网络仍然可以产生相当准确的预测，从而为实际湍流系统的无模型预测铺平了道路。
In the problem ${\rm Mix}(H)$ one is given a graph $G$ and must decide if the Hom-graph ${\rm {\bf Hom}}(G,H)$ is connected. We show that if $H$ is a triangle-free reflexive graph with at least one cycle, ${\rm Mix}(H)$ is ${\rm coNP}$-complete. The main part of this is a reduction to the problem ${\rm NonFlat}({\rm{\bf H}})$ for a simplicial complex ${\rm{\bf H}}$, in which one is given a simplicial complex ${\rm{\bf G}}$ and must decide if there are any simplicial maps $ϕ$ from ${\rm{\bf G}}$ to ${\rm{\bf H}}$ under which some $1$-cycles of ${\rm{\bf G}}$ maps to homologically non-trivial cycle of ${\rm{\bf H}}$. We show that for any reflexive graph $H$, if the clique complex ${\rm{\bf H}}$ of $H$ has a free, non-trivial homology group $H_1({\rm{\bf H}})$, then ${\rm NonFlat}({\rm{\bf H}})$ is ${\rm NP}$-complete.
