Numerical null controllability of semi-linear 1-D heat equations: fixed point, least squares and Newton methods (Q1760182)

This paper deals with the analysis and the implementation of numerical methods allowing to solve the null controllability problem for the following semi-linear one-dimensional heat equation: \[ y_t-(D(x)y_x)_x+f(y)=v1_\omega,\,(x,t)\in (0,1)\times (0,T), \] \[ y(x,t)=0,\,(x,t)\in \{0,1\}\times (0,T); \quad y(x,0)=y_0(x),\,x\in (0,T), \] where \(T>0,\) \(\omega \subset\subset (0,1)\) is a non empty open interval, \(v\in L^\infty (\omega\times (0,T))\) is the control, \(y\) is the associated state and \(1_\omega\) is the characteristic function of the control set. The function \(D\) belongs to \(C^1([0,1])\) with \(D(x)\geq D_0>0\) and \(y_0\in L^2(0,1)\). The function \(f:\mathbb R\to\mathbb R\) is supposed to be at least locally Lipschitz-continuous. The goal is achieved by means of standard fixed point formulation, gradient techniques and Newton-Raphson strategy. Some data for which least squares algorithm converge are exhibited.