No label defined (Q4526909)

The authors study the following inverse problem for a reaction-diffusion system, that is, for determining the four functions \((r,s,D,k)\) satisfying the following system of equations: \[ \begin{aligned} r_t-D(x)r_{xx}&=kf_1(r)s,\quad x\in (0,1),\quad t\in (0,T),\\ r s_t&=kf_2(r)s, \quad s(x,0)=1, \\ r(0,t)&=r(1,t)=0, \quad r(x,0)=1,\\ r(x,T)&=z(x), \quad g(t)=\int_{0}^{1} r(x,t) dx, \end{aligned} \] where \(f_1, f_2, g\), and \(z\) are given and \(k\) is constant. First, they approximate \(D(x)\) using a finite element method, and then deduce the above inverse problem to an optimization problem for a nonlinear functional with respect to the parameter \(D\). They prove that the previous optimization problem has a solution. Finally, they conclude that the unknown parameters \(D(x)\) and \(k\) can be uniquely determined by the data \(z(x)\) and \(g(t)\).