Dirichlet boundary control of semilinear parabolic equations. I: Problems with no state constraints (Q1599470)

The system is \[ \begin{aligned} {\partial y(t, x) \over \partial t} + Ay(t, x) + \Phi(t, x, y(t, x), u(t, x)) = 0 \quad &(0 \leq t \leq T, \;x \in \Omega) \\ y(0, x) = w(x) \quad (x \in \Omega) , \qquad y(t, x) = v(t, x) \quad &(0 \leq t \leq T, \;x \in \Gamma) \end{aligned} \tag{1} \] where \(\Omega\) is an \(m\)-dimensional domain with boundary \(\Gamma\) and \(A\) is a second order elliptic operator. The controls in (1) are \(u(t, x)\) in \(L^q((0, T) \times \Omega),\) \(v(t, x)\) in \(L^\infty((0, T) \times \Gamma)\) and the initial condition \(w(x)\) in \(L^\infty(\Omega),\) all three subject to control constraints. The authors prove decoupled Pontryagin's maximum principles for each of the three controls. Due to the presence of the \(L^\infty\) boundary control and initial condition, solutions of the state equation take values in the space \(C_b((0, T] \times \Omega)\) of bounded continuous functions in \((0, T] \times \Omega.\) To place solutions in this space makes it possible to handle pointwise constraints on the state, which the authors announce for Part II of this work [ibid. 45, No. 2, 145--167 (2002; Zbl 1037.49023)].