Controllability of a class of Volterra equations in Hilbert spaces with completely monotone kernel (Q692800): Difference between revisions

The authors consider the linear Volterra integro-differential equation with memory in a Hilbert space \(H\): \( \frac{d}{dt}\int_{-\infty}^t a(t-s)u(s)~ds=Au(t)+\Gamma y(t), \quad t\in (0,T];\) \( u(s)=u_0(s)=\varphi(s), \quad s\leq 0\). Here the operator \((-A): D(A)\subset H\rightarrow H\) is sectorial, the kernel \(a:(0,\infty)\rightarrow (0,\infty)\) is completely monotone, with \(a(0+)=+\infty\), and the linear operator \(\Gamma:U\rightarrow H\) is bounded from some Hilbert space \(U\) into \(H\). The admissible control function \(y\) is in \(L^2(0,T;U)\) and \(\varphi:(-\infty,0)\rightarrow H\) is a suitable \(H\)-valued function. The state of the system, \(v(t)\), which contains all information about the solution up to time \(t\) and the solution \(u(t)\), are governed by the following abstract differential equation in a larger Hilbert space \(X\): \( v'(t)=Bv(t)+(I-B)P\Gamma y(t); ~~ v(0)=v_0;\) \(u(t)=Jv(t)\), where \(B:D(B)\subset X\rightarrow X\) is the infinitesimal generator of an analytic semigroup, \(P: H\rightarrow H\) is a bounded linear operator, \(J:D(J)\rightarrow H\) is an unbounded operator. The authors show that, for a large class of kernels, the state cannot be controlled exactly to zero. However, equations of the considered type are always approximately controllable, provided the control operator has dense range.