Uniqueness of the solution of some inverse problems in nonequilibrium sorption dynamics (Q1086437)

The author considers the following problem occurring in the dynamics of sorption: \[ u_ x+a_ t=0\quad in\quad \Omega =(0,+\infty)\times (0,T),\quad T\in (0,+\infty) \] \[ a_ t=f(u)-g(a)\quad in\quad \Omega;\quad u(0,t)=h(t)\quad 0\leq t\leq T \] \[ a(x,0)=k(x)\quad 0\leq x;\quad u(\ell,t)=m(t)\quad 0\leq t\leq T,\quad \ell >0. \] Such a problem is obviously overdetermined, when all the functions f,g,h,k,m are supposed to be known. On the contrary the author assumes that some of them are unknown and considers the related identification problem. More specifically the three following problems are investigated: (I.P.1) to determine the sextuplet (u,a,\(\beta\),\(\gamma\),h,k), when m is assumed to be known and the functions f and g are of the following forms: f(s)\(\equiv \beta s\), g(s)\(\equiv \gamma s\), \(\beta\) and \(\gamma\) being unknown positive constants; (I.P.2) to determine the triplet (u,a,f), when m is assumed to be known and g(s)\(\equiv s\), k(x)\(\equiv 0;\) (I.P.3) to determine the triplet (u,a,g), when m is assumed to be known and f(s)\(\equiv s\), k(x)\(\equiv 0.\) The author proves the following three results concerning local uniqueness: (R.1) suppose that \(T=+\infty\) and \(h\in C([0,+\infty))\setminus \{0\}\), h(t)\(\geq 0\forall t\in [0,t_ 0]\) for some positive \(t_ 0\) and \(h(t)=0\forall t\in [t_ 0,+\infty)\). If \((u_ j,a_ j,\beta_ j,\gamma_ j,h_ j,k_ j)\) \((j=1,2)\) are solutions to the identification problem (I.P.1), then \(\beta_ 1=\beta_ 2\), \(\gamma_ 1=\gamma_ 2\), \(h_ 1(t)=h_ 2(t)\forall t\in [0,+\infty)\), \(k_ 1(x)=k_ 2(x)\forall x\in [0,\ell]\), \(u_ 1=u_ 2\) and \(a_ 1=a_ 2\) in \([0,\ell]\times [0,+\infty)\); (R.2) Suppose that \(T<+\infty\) and h is strictly positive and non-decreasing and non-decreasing in [0,T]. Assume moreover that \((u_ j,a_ j,f_ j)\) \((j=1,2)\) are solutions to problem (I.P.2), \(f_ j\in C^ 1(R)\) \((j=1,2)\) being real-analytic functions in an open interval containing [0,h(T)], such that \(f_ j(0)=0\) and \(f_ j'(s)>0\forall s\in R\). Then \(u_ 1=u_ 2\) and \(a_ 1=a_ 2\) in [0,\(\ell]\times [0,T]\) and \(f_ 1(s)=f_ 2(s)\forall s\in [0,h(T)]\). (R.3) Under assumptions similar to the previous ones (with \(f_ j\) and h(T) replaced respectively by \(g_ j\) and \(f^{-1}(h(T)))\) we have \(u_ 1=u_ 2\) and \(a_ 1=a_ 2\) in [0,\(\ell]\times [0,T]\) and \(g_ 1(s)=g_ 2(s)\forall s\in [0,f^{- 1}(h(T))]\).