论文标题

顶点分区和最大$ \ g $ - 免费子图

Vertex Partitions and Maximum $\G$-free Subgraphs

论文作者

Rowshan, Yaser

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。

We define a $(V_1, V_2, \ldots, V_k)$-partition for a given graph $H$ and graphical properties $P_1, P_2, \ldots, P_k$ as a partition where each $V_i$ induces a subgraph of $H$ with property $P_i$. Matamala (2007) extended this result by showing that for any graph $H$ with $Δ(H)=p+q$, there exists a $(V_1, V_2)$-partition of $V(H)$ where $H[V_1]$ is a maximum order $(p-1)$-degenerate induced subgraph and $H[V_2]$ is $(q-1)$-degenerate. Additionally, Catlin and Lai proved that if $Δ(H)\geq 5$, $H$ has a $(V_1, V_2)$-partition such that $H[V_1]$ is a maximum order acyclic induced subgraph, $ω(H[V_2])\leq Δ(H)-2$, and $Δ(H[V_2])\leq Δ(H)-2$. Rowshan and Taherkhani demonstrated that given a graph $G$ with a minimum degree $δ(G)$ and for $k=\lceil \frac{Δ(H)}{δ(G)}\rceil$, there exists a $(V_1, V_2, \ldots, V_k)$-partition of the vertex set of $H$, such that each $H[V_i]$ is $G$-free, meaning it does not contain a subgraph isomorphic to $G$, and $H[V_1]$ is a maximum order $G$-free induced subgraph. In our paper, we present a novel result for a connected graph $H$ with $Δ(H)\geq 5$ and without $K_{Δ(H)+1}\setminus e$ as a subgraph. We establish that when $p_1\geq p_2\geq\cdots\geq p_{k-1}\geq 2$, $p_k\geq 4$, $\sum_{i=1}^k p_i=Δ(H)-1+k$, and $\mathcal{G}_i$ represents a family of graphs with a minimum degree at least $p_i-1$ for each $i\in [k-1]$, a $(V_1, V_2, \ldots, V_k)$-partition of $V(H)$ exists. This partition guarantees that $H[V_1]$ is a maximum order $\mathcal{G}_1$-free induced subgraph, $H[V_i]$ is $\mathcal{G}_i$-free for each $2\leq i\leq k-1$, $Δ(H[V_k])\leq p_k$, and either $H[V_k]$ is $K_{p_k}$-free or its $p_k$-cliques are disjoint.

扫码加入交流群

加入微信交流群

微信交流群二维码

扫码加入学术交流群,获取更多资源