Computers in Physics
Index of content:
Volume 5, Issue 1, March 1991
- PEER-REVIEWED PAPERS
5(1991); http://dx.doi.org/10.1063/1.168405View Description Hide Description
A procedure has been developed to calculate numerically the envelope functions of an oscillatory curve. The method has been shown to be applicable to optical transmission data, but it is general enough to be used for many other data sets. The program is available on request.
5(1991); http://dx.doi.org/10.1063/1.168406View Description Hide Description
The phenomena associated with the reflection and refraction of light have been widely studied. As the level of dimensionality of the model used in the analysis increases, so does the variety of interesting phenomena that emerge from the analysis. For example, plane-wave calculations give rise to Brewster angle and total internal reflection effects. If a finite two-dimensional slab beam is considered, Goos-Hänchen, focal, and angular shifts become apparent. As the problem is generalized to higher dimensions, possibly including the temporal dimension, additional phenomena may be expected as well. This paper begins by defining the terms and general notation used with the interface problem by reviewing plane-wave (infinite field) solutions. This is followed by a literature review of some two-dimensional results, including a discussion of the Goos-Hänchen, focal, and angular shift. Then numerical solutions to the full three-dimensional problem of finite beams (under a paraxial approximation) are presented using modern visualization techniques. The numerical results provide an intuitive understanding of the interface phenomena from a new perspective, graphically underscoring the difference between the geometrical ray model of reflection and refraction and the complicated field interaction that actually occurs. Finally, the numerical model is extended to include results for the nonlinear interface.
5(1991); http://dx.doi.org/10.1063/1.168431View Description Hide Description
Although much has been written about relativity, little delves into its actual equations. Even rarer does one find relativity used in any sort of computation, such as one would find in a computer simulation. This is sad. it is hoped to be shown here not just the problems one might encounter in writing such a simulation, but also that one can perform a simulation of general relativity without too much effort. In fact, our experience has been that a program simulating geodesic motion in general relativity requires no more lines of code than a corresponding program to simulate planetary motion using Newton's laws. The relativity program is straightforward and general, but slow. However, by mechanically substituting the equations for the gravitational time-dilation metric of the Schwarzchild metric, and then performing algebraic simplifications, one can obtain a program whose simulation time is as fast as a corresponding Newtonian motion simulation. The resulting customized relativity programs are again short and simple. Customization does, however, obscure a little relation between the general equations and those actually used.
5(1991); http://dx.doi.org/10.1063/1.168418View Description Hide Description
In this work, a parallel signal processing environment will be introduced. An initial testing system has been implemented and a specific parallel hardware and software are under consideration. An underwater acoustic lattice code is used as a case study in the prototype environment. The experiments prove that the parallel signal processing environment suggested will be a powerful tool for both physicists and engineers. The major improvement of the approach is the combination of a parallelism and a neural network pattern recognition. Another innovation is use of a dynamic lattice code instead of the fixed-size lattice code.