The multi-resolution method applied to the sideways heat equation (Q2566791)

The author considers the problem of finding the function \(u(x,t)\) from \(u_{xx}(x,t) = u_t(x,t), 0 \leq x < 1, t \geq 0, u(1,t) = g(t), t \geq 0\). He calls this problem ``the sideways heat equation''. As the problem is ill-posed, the author uses the multi-resolution method based on the Shannon multi-resolution analysis (MRA) to obtain a well-posed approximating problem and then an error estimate between the exact solution and the solution of the approximating problem defined in a subspace of the MRA. The problem setting is wrong and Theorem 1.1 in the paper is not correct. Furthermore, (1) in its setting is not the sideways heat equation. The heat equation \(u_{xx}(x,t) = u_t(x,t)\) has to be considered in the quarter \((x,t) \in (0,\infty)\times (0,\infty)\) rather than in \((0,1) \times (0,\infty)\). Furthermore, some bound should be imposed on the solution at infinity. A more general problem has been theoretically and numerically studied in the reviewer's and his co-workers' papers [Inverse Probl. 13, No.2, 297--309 (1997; Zbl 0871.35105)], [Inverse Probl. 11, No.6, 1247-1263 (1995; Zbl 0845.35130); J. Math. Anal. Appl. 199, No.3, 873-909 (1996; Zbl 0861.35033); Int. J. Numer. Methods Eng. 50, No.5, 1253--1267 (2001; Zbl 1082.80003)].