The Hamiltonian operator associated with some quantum stochastic evolutions (Q5953672)

Consider the quantum stochastic differential equation introduced by Hudson and Pathasarathy \[ dV_t = \left[ \sum_{i,j\in J} (S_{i,j}-\delta_{ij})d\Lambda_{ij}(t) - \sum_{i,j\in J} R^*_iS_{ij}dA_j(t) + \sum_{j\in I}R_jdA^{\dag}_j(t) \right. \] \[ \left. - \left( iH+{1\over 2} \sum_{j\in J} R^*_j R_j \right) dt \right] V_t, \] \(V_0 = {\mathbf 1}\), where \(S_{ij},R_j,H\) are operators on a Hilbert space \({\mathcal H}\), and \(d\Lambda_{ij}(t)\), \(dA_j(t)\), \(dA^{\dag}_j(t)\), \(i,j\in J\), are the basic quantum stochastic differentials on the symmetric Fock space \(\Gamma (L^2(R_+;\zeta))\) with arbitrary multiplicity \(J\), i.e. \(\zeta\) is identified with \(l^2(J)\) and \(J\) is an at most countable set. The solution \((V_t)_{t\in R_+}\) of this equation is an adapted process of operators on \(\Gamma (L^2(R_+;\zeta))\otimes {\mathcal H}\) that describes the evolution of an open system on a Hilbert space \({\mathcal H}\), coupled to a boson field on the symmetric Fock space \(\Gamma (L^2(R_+;\zeta))\). Let \(\Theta_t\), \(t\in R_+\), denote the second quantization of the shift operator on \(L^2(R;\zeta)\). Given the cocycle property \(V_{s+t} = \Theta_s^* V_t \Theta_s V_s\), a strongly continuous one-parameter unitary group of operators \((U_t)_{t\in R}\) on \(\Gamma (L^2(R;\zeta)) \otimes {\mathcal H}\) is constructed as \[ U_t = \Theta_t V_t, \;\;\;\;\;U_{-t} = \Theta_{-t}^* V_t, \;\;\;t\geq 0. \] Let \({\mathbf H}\) denote the Hamiltonian associated to this equation, i.e. \({\mathbf H}\) is the generator of \((U_t)_{t\in R}\). In a previous paper [\textit{M. Gregoratti}, Infin. Dimens. Anal. Quantum Probab. Relat. Top. 3,483-503 (2000; Zbl 1042.81055)], and in the case of single multiplicity, the author has constructed a densely defined restriction of \({\mathbf H}\) on a domain of vectors \(\sum_{n=0}^\infty \Phi_n \in \Gamma (L^2(R; \zeta))\otimes {\mathcal H}\) whose kernels \(\Phi_n\) are Sobolev differentiable, and proved in some cases (e.g. when \(R=0\)) the self-adjointness of this restriction. This paper treats the case of arbitrary multiplicity and extends previous results of the author by finding an essentially self-adjoint restriction of \({\mathbf H}\), under boundedness assumptions on the system operators \(H,R_j,S_{i,j}\).