[en] This is a short book of 147 pages including references and sometimes bibliographies at the end of each chapter, and subject and author indices at the end of the book. The test includes an introduction of 3 pages, 29 pages explaining approximate analysis, 41 pages on finite differences, 36 pages on finite elements, and 17 pages on specialized methods

[en] At present, the Amplitude-Magnitude (AM) method is largely applied for determining the spin axis direction of asteroids; such a method gives results which in general are in a good agreement with those obtained by means of fundamentally different methods. However, one of the critical assumptions of the AM method is that asteroids are modelled as triaxial ellipsoids with semiaxes a>b>c. Although such an hypothesis appears reasonable for large objects on the basis of physical considerations about their expected equilibrium shapes, very irregular figures are more plausible for smaller objects, probably dominated by solid-state forces. Therefore, it is worth-while to study the influence that deviations from a purely triaxial ellipsoid shape can have on the derived spin axis direction. For this purpose, a numerical program has been developed in order to compute the light curves of irregularly shaped objects at given aspect angles. Some preliminary results are reported concerning the uncertainty of the rotation axis direction of asteroids when non-triaxial shapes are considered. (author). 4 figs., 1 tab., 8 refs

[en] RF-driven underdamped Josephson junctions are expected to exhibit chaotic behavior in the rf frequency range roughly defined by ω_RC< or approx.ω< or approx.ω_p where ω_RC=1/RC and ω_p=(2cI_c/ℎC)^1/2. Numerical simulations of the I-V curves are presented as the rf frequency is increased from 0.1 ω_RC to 2 ω_p. The effect of chaos on the I-V curves is strongly dependent on the rf frequency. (orig.)

No abstract available

[en] In this paper a BEM is used to solve a class of variable coefficient elliptic equations numerically. Some examples are considered to show the convergence, consistency, and accuracy of the numerical solutions. (paper)

No abstract available

[en] For finding multiple solutions of simultaneous nonlinear equations, we present a homotopy method, based on integrating a regular differential equations. This differential system is solved by a predictor-corrector scheme, based on Hermite corrector formula; and we obtain the approximate solutions of the system of nonlinear equations. Then, with the approximate solutions as initial points, the method of Newton-Raphson gives multiple solutions of the system of nonlinear equations

[en] Automatic differentiation is a method of computing derivatives of functions to any order in any number of variables. The functions must be expressible as combinations of elementary functions. When evaluated at specific numerical points, the derivatives have no truncation error and are automatically found. The method is illustrated by simple examples. Source code in FORTRAN is provided

[en] A perturbative numerical algorithm is developed for the solution of the one-dimensional Schroedinger equation. This starts with the step-function reference potential for which analytic solution and its first derivative are calculated, and then first and second order perturbative corrections are added. The present algorithm shows the following features: (i) its order of convergence is O(hsup(7 approximately 9)) depending on the expression of the potential function; (ii) it reaches high accuracy even for very coarse partitions of the integration domain; (iii) it is free of computational difficulties such as the subtractive near-cancellation effects. A numerical example illustrates the high efficiency of the present algorithm. (author)

No abstract available