Diffusion subject to the specification of mass (Q1081751): Difference between revisions

The paper demonstrates that the problem \[ u_ t=u_{xx}+s(x,t),\quad 0<x<1,\quad 0<t\leq T, \] \[ u(x,0)=f(x),\quad 0<x<1;\quad u_ x(1,t)=g(x),\quad 0<t\leq T, \] \[ m(t)=\int^{b(t)}_{0}u(x,t)dx,\quad 0<b(t)<1,\quad 0<t\leq T, \] where b,f,g,m and s are known functions, T is a given positive constant, and \(m(0)=\int^{b(0)}_{0}f(x)dx\), possesses a unique solution that depends continuously upon the data. In section 2 an equivalent Volterra integral equation of the second kind is derived, while in section 3 the existence, uniqueness and continuous dependence are proved. Also, as a corollary of the proposed method a simpler integral equation is derived in section 4, for the semi-infinite case \[ u_ t=u_{xx}+s(x,t),\quad 0<x<\infty,\quad 0<t\leq T;\quad u(x,0)=f(x),\quad 0<x<\infty, \] \[ m(t)=\int^{b(t)}_{0}u(x,t)dx,\quad 0<b(t)<\infty,\quad 0<t\leq T, \] where \(m(0)=\int^{b(0)}_{0}f(x)dx.\) Finally, in section 5 a numerical method is presented with a discussion of its application to the quarter-plane case and several numerical results are also given. The paper is very clearly written and it is also very interesting because, the problem being examined, can model several physical problems.