Diffusion of power in randomly perturbed Hamiltonian partial differential equations (Q1781826)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Diffusion of power in randomly perturbed Hamiltonian partial differential equations
scientific article

    Statements

    Diffusion of power in randomly perturbed Hamiltonian partial differential equations (English)
    0 references
    0 references
    0 references
    8 June 2005
    0 references
    The authors study the system: \(i\partial_t\phi(t)= (H_0+\varepsilon g(t)\beta)\phi(t, x)\), where \(H_0\) and \(\beta\) are self-adjoint operators on the Hilbert space \({\mathcal H}\). Let \(g_0(t)\) be a fixed real-valued function with support in \([0,T]\), and \(\{d_j\geq 0\}_{j\geq0}\) a random sequence. They treat the systems with \(g(t)= \sum_{n=0\sim\infty}g_0(t- t_n)\); \(t_0= d_0\), \(t_n= d_n+ \sum_{i= 0\sim(n-1)})d_i+ T)\), \(n\geq 1\). (I) Suppose that \(H_0\) has only a discrete spectrum, \(\|\beta\|= 1\), and \(g= g_0\) with \(\| g_0\|_1= 1\). That is, \(H_0\psi_j= \lambda_j\psi_j\), \(\langle\psi_i,\psi_j\rangle= \delta_{ij}\). Assume that \(\{a_j(t)\}\) in \(\phi(x, t)= \sum_j a_j(t)\psi_j(x)\) satisfies \(E(a_j(0)\overline{a_k(0)})= 0\) for \(j\neq k\). Then \[ \begin{multlined} P_k(t)(:= E(|a_k(t)|^2))= P_k(0)+ \varepsilon^2\sum_j|\langle\psi_j, \beta\cdot\psi)\rangle|^2\cdot\\|{\mathcal F}g_0(\lambda_j- \lambda_k)|^2(P_j(0)- P_k(0))+ O(\varepsilon^3)\equiv (T_\varepsilon P(0))_k+ O(\varepsilon^3),\end{multlined} \] holds for \(t> T\). For the system with \(\{d_j\}\) satisfying \(E(\exp(i(\lambda_j- \lambda_k)d_m)= 0\) for \(j\neq k\), \[ P^{(n)}(:= \{P_k^{(n)}\equiv E(|a_k(t)|^2)\})= (T_\varepsilon)^n P(0)+ O(n\varepsilon^3), \] \(t_{n-1}+ T\leq t\leq t_n\), holds for \(n\geq 1\). (II) The spectrum of \(H_0\) has an absolutely continuous part: \[ \phi(t, x)= \sum_{j= 1\sim m} a_j(t)\psi_j(x)+ P_c\phi(x,t). \] Then \(P(t)= (T_{\text{disc},\varepsilon}- \varepsilon^2\Gamma)^n P(0)+ O(\varepsilon^{4/5})\), holds for \(t> 0\). Here, \(n\) is the integer part of \(t/(T+ M)\), and \(M\) is the mean of \(\{d_j\}\).
    0 references
    energy distribution
    0 references
    discrete modes
    0 references
    discrete and continuous spectrum
    0 references
    discrete diffusion-damping
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references