A stable space marching finite differences algorithm for the inverse heat conduction problem with no initial filtering procedure (Q914345)

The authors consider the one-dimensional semi-infinite linear inverse sideways heat conduction problem and present a new algorithm based on singular perturbation techniques. The concept of formal stability for this problem is analyzed and its relationship with the high frequency components of the measured data function is studied by comparing the new method with the mollification method and the method of \textit{C. F. Weber} [Int. J. Heat Mass Transfer 24, 1783-1792 (1981; Zbl 0468.76086)]. Finally, a fully explicit space marching finite differences scheme is developed and a numerical example is presented showing the accuracy and stability of the algorithm in the presence of noisy data and for small dimensionless timesteps. The reviewer wants to mention that for the same numerical example a similar result was achieved by Tikhonov regularization with optimal parameter choice applied to the equivalent integral equation representation of the sideways heat equation [\textit{J. T. King} and the reviewer, Computing 40, 91-109 (1988; Zbl 0629.65138)].