The utilization of total mass to determine the switching points in the symmetric boundary control problem with a linear reaction term (Q2568951): Difference between revisions

The paper studies the initial-boundary value problem for a heat equation of one dimension. The equation has state \(u(x, t)\), and is described by \[ u_t= u_{xx}- u,\quad 0< x< 1,\quad t> 0, \] \[ u(x,0)= 0,\quad 0< x< 1,\quad\text{and }u(0,t)= u(1,t)= \psi(t),\quad t> 0, \] where the control \(\psi(t)\) is such that \(\psi(t)= 1\) on the interval \((t_{2k}, t_{2k+1})\); \(\psi(t)= 0\) on \((t_{2k+1}, t_{2k+2})\); \(\mu(t_{2k})= m\); and \(\mu(t_{2k+1})= M\), \(\mu(t)\) denoting the total mass: \(\int^1_0 u(x, t)\,dx\). The problem is to determine the increasing sequence \(\{t_n\}\) of positive numbers (switching points) and the numbers \(0< m< M\): When \(\psi(t)\) is replaced by the unit function \(1\), \(t\geq 0\), \(\mu(t)\) is simply increasing: \(0< \mu(t)< \mu_\infty= 2(e-1)/(e+ 1)\). Thus \(t_1\) with \(\mu(t_1)= M\) is uniquely determined for \(0<\forall M< \mu_\infty\). Next replace \(\psi(t)\) by \(1\) on \(0\leq t< t_1\) and \(0\) on \(t> t_1\). Then \(\mu(t)\) is simply decreasing: \(M> \mu(t)> 0\). Thus \(t_2\) \((> t_1)\) with \(\mu(t_2)= m\) is uniquely determined for \(0< \forall m< M\). The whole sequence \(\{t_n\}\) is obtained by repeating the process in this manner. An approximate estimate of the differences: \(t_k- t_{k-1}\), \(k= 1,2,\dots\) is also obtained.