No label defined (Q4353689)

The system is described by the equation \[ {\partial y\over\partial t}+ Ay+ \Phi(x,t,y)= 0\quad (0\leq t\leq T,\;x\in\Omega=\text{bounded domain in }\mathbb{R}^N), \] where \(A\) is a second-order symmetric uniformly elliptic operator with regular coefficients. The control \(v\) appears in the boundary condition: \[ {\partial y\over\partial n_A}+ \Psi(y,v)= 0\quad (0\leq t\leq T,\;x\in\Gamma=\text{boundary of }\Omega). \] The control constraints are: \(v\in\)closed, convex set \(K\subseteq L^r([0,T]\times \Omega)\) for suitable \(r\). There are also state constraints of the form \[ g(x,t,y(t,x))\leq 0 \] and the cost functional to be minimized is \[ J(y,v)= \int_{[0, T]\times\Omega} F(x,t,y)dx dt+ \int_{[0, T]\times\Gamma} G(y,v)d\sigma dt+ \int_\Omega H(x,y(x, T))dx. \] Under suitable assumptions on the domain and the various functions and coefficients, the author obtains a version of Pontryagin's maximum principle for solutions of this optimization problem where, due to the pointwise state constraints, the adjoint vector satisfies a parabolic equation with measures as source terms. This paper is an announcement of results to be published in full elsewhere. For the entire collection see [Zbl 0865.00062].