The generalized eigenproblem: pole-zero computation
Article Abstract:
In linear system analysis the eigenproblem is one of the fundamental problems. Normally, direct application of the QZ algorithm is used to solve network eigenvalue problems. Here Haley introduces a simple, reliable, fundamental method based on matrix modification for solving a general class of eigenproblems. Called the modification-decomposition (MD) method, it computes linear system transfer function poles and zeros by transforming an N-dimensional generalized eigenvalue problem. A general eigenvalue sensitivity formula is derived, and a flexible eigenvector computation method is developed and applied to pole sensitivity computation. Haley expects that a variation of the MD method will find wide use in the solution of general eigenproblems. It can, for example, be directly implemented in any nodal based circuit simulator, such as SPICE, by adding a QR eigenvalue routine.
Publication Name: Proceedings of the IEEE
Subject: Electronics
ISSN: 0018-9219
Year: 1988
User Contributions:
Comment about this article or add new information about this topic:
Visual shape computation
Article Abstract:
Research in computer vision has shifted from domain dependent imaging to modular vision systems corresponding to, in most cases, operations in human visual systems, such as the extraction of shape from shading, texture, contour, motion and stereo. Algorithms for the determination of shape should utilize an active vision approach, wherein active observations of the trajectories of an object in stimulus space constitutes an additional source of shape information. Consequently, shape from shading, contour, texture and motion constitute well-posed problems with unique solutions. These processes, except structure from motion, are ill-posed problems in a passive vision environment. Details of the shape extraction processes are described.
Publication Name: Proceedings of the IEEE
Subject: Electronics
ISSN: 0018-9219
Year: 1988
User Contributions:
Comment about this article or add new information about this topic:
A heat engine model of a reversible computation
Article Abstract:
The tools of information theory and thermodynamics can be used to model reversible computation at the binary level as a heat engine operating in a Carnot cycle. The binary bits are energized with a characteristic energy. The engineer can compute the derivative of information-theoretic entropy with respect to energy for a binary bit stream, resulting in the concept of information-theoretic temperature as a fundamental metric. This, in turn, allows the Boltzmann factor to be used to describe the statistical properties of the bit stream. The heat engine model is consistent with existing models of computation.
Publication Name: Proceedings of the IEEE
Subject: Electronics
ISSN: 0018-9219
Year: 1990
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: Dynamic algorithms in computational geometry. Parallel techniques for computational geometry
- Abstracts: Man-machine interface problems in designing air traffic control systems. Transient mean and variance of overflow traffic in a loss system
- Abstracts: Runlength-limited sequences. On the computation of motion from sequences of images - a review