由Dirichlet拉普拉斯算子多项式生成的球约束流中的谱选择
Spectral Selection in Sphere-Constrained Flows Generated by Polynomials of the Dirichlet Laplacian
浏览论文内容
中文总结 AI 辅助
该研究针对有界光滑区域上Dirichlet拉普拉斯算子多项式生成的球约束流，分析其轨道收敛性、谱选择规律及扰动稳定性，明确不同多项式对应的谱选择机制。
中文摘要 AI 辅助
设Ω⊂ℝᵈ是有界光滑区域，-Δ_D是L²(Ω)上的正Dirichlet拉普拉斯算子。对于首项系数为正的实多项式p，我们研究约束线性方程：u_t = -Πᵤ p(-Δ_D)u，||u(0)||_{L²}=1，其解为归一化半群轨道：u(t)=e^{-t p(-Δ_D)}u₀ / ||e^{-t p(-Δ_D)}u₀||_{L²}。该轨道的主动谱支撑保持不变，且收敛到u₀在主动特征子空间上的归一化投影，其中p(λⱼ)最小；下一个主动多项式谱值给出指数速率。对任意θ≥0和τ>0，即使初始数据无分数阶正则性，在(-Δ_D)^θ的定义域内t≥τ时仍保持相同速率。我们还证明，可将有限个Dirichlet能级指定为多项式的全局极小集，且孤立选择集在足够小的多项式扰动下稳定。当p(s)=sᵐ时，选择最低主动Dirichlet能级；当p(s)=(s-ρ)²时，按到ρ的距离选择，跨能级简并仅发生在Dirichlet中点处。
英文摘要
Let $Ω\subset\mathbb{R}^d$ be a bounded smooth domain and let $-Δ_D$ be the positive Dirichlet Laplacian on $L^2(Ω)$. For a real polynomial $p$ with positive leading coefficient, we study the constrained linear equation \[ u_t=-Π_u\,p(-Δ_D)u, \qquad \lVert u(0)\rVert_{L^2}=1. \] Its solution is the normalized semigroup orbit \[ u(t)=\frac{e^{-t p(-Δ_D)}u_0} {\lVert e^{-t p(-Δ_D)}u_0\rVert_{L^2}}. \] The active spectral support is preserved, and the trajectory converges to the normalized projection of $u_0$ onto the active eigenspaces for which $p(λ_j)$ is minimal. The next active polynomial spectral value gives the exponential rate. For every $θ\geq 0$ and $τ>0$, the same rate holds in the domain of $(-Δ_D)^θ$ for $t\geqτ$, even when the initial datum has no fractional regularity. We also show that a finite set of Dirichlet levels can be prescribed as the global minimizing set of a polynomial, and that an isolated selected set is stable under sufficiently small polynomial perturbations. For $p(s)=s^m$, the lowest active Dirichlet level is selected. For $p(s)=(s-ρ)^2$, selection is by distance from $ρ$, and cross-level degeneracy occurs only at Dirichlet midpoints.
