Optimal design of large software-systems using N-versionprogramming

Fault tolerant software uses redundancy to improve reliability; but such redundancy requires additional resources and tends to be costly, therefore the redundancy level needs to be optimized. Our optimization models determine the optimal level of redundancy within a software system under the assumption that functionally equivalent software components fail independently. A framework illustrates the tradeoff between the cost of using N-version programming and the improved reliability for a software system. The 2 models deal with: a single task, and multitask software. These software systems consist of several modules where each module performs a subtask and, by sequential execution of modules, a major task is performed. Major assumptions are: 1) several versions of each module, each with an estimated cost and reliability, are available, 2) these module versions fail independently. Optimization models are used to select the optimal set of versions for each module such that the system reliability is maximized and total cost remains within budget     

Broad-band array design using the asymptotic theory of unequallyspaced arrays

The problem of designing broadband linear arrays of omnidirectional sensors is addressed. Attention is restricted to the case of shaded delay-and-sum beamforming. Broadbanding is here defined to mean that the beam-pattern function has little or no frequency dependence of peak response, main-lobe width, plateau sidelobe level, and sinespace separation between the main lobe and the plateau sidelobe. The asymptotic theory of unequally spaced arrays is used to derive relationships between beam-pattern properties and array properties. These relationships are used to translate beam-pattern requirements into functional requirements on sensor spacings and amplitude shadings. The functional requirements are then used to derive a broadband array design. In addition to the design equations, the asymptotic theory is used to derive equations for main-lobe level, sidelobe level, beamwidth, and signal-to-noise ratio (SNR) gain in isotropic noise. A specific example is presented to clarify the concepts and verify that the design procedure actually works     

Optimal windowing for wideband linear arrays

The problem of constructing a wideband beam pattern that is concentrated in both angle and frequency is discussed. This paper is a direct extension of the work of Slepian and co-workers on time- and frequency-limited functions. It is shown that the singular vectors and singular functions of the mapping relating the set of weights of a linear wideband array to its far-field directivity pattern have both concentration properties and double orthogonality properties and so they can be thought of as the wideband equivalents of the discrete prolate spheroidal sequences and wave functions. These singular functions are used to obtain approximations to a frequency-invariant beam pattern.     

Optimal design for micro-thermoelectric generators using finite element analysis

To fabricate micro-thermoelectric generators (μTEGs), one must design the optimal structure of the μTEGs and achieve thermoelectric thin films with excellent properties. This study investigated the role of the dimensions of μTEGs, including the length of the thermoelements, thickness of the substrates, and cross-sectional area of the thermoelements. To evaluate the power generated by μTEGs and their efficiency, three-dimensional models of μTEGs were subjected to finite element analysis. Three-dimensional models are more accurate than one-dimensional models, since the directions of the heat flux and electrical current are not parallel in μTEGs. The governing equations were derived from the Seebeck effect and Peltier effect, which show thermoelectric energy conversion. In the simulation, the substrate, n-type material, and p-type material were assumed to be silicon, Bi₂Te₃, and Sb₂Te₃, respectively. We calculated the thermoelectric power generated by the μTEGs and their thermoelectric energy conversion efficiency. These two evaluation indices represent the performance of μTEGs. The thermoelectric simulation produced design guidelines for high-performance μTEGs.     

Optimal design for broadband quasi-phase-matched second-harmonicgeneration using simulated annealing

We propose a modulation-optimization method for achieving broadband quasi-phase matching in second-harmonic generation devices with domain-inverted structures. Using simulated annealing, we optimize phase-matching curves tailored to practical purposes by appropriately choosing a cost function in the annealing process. Both the acceptance bandwidth and the conversion efficiency are significantly improved in comparison with conventional periodically domain-inverted devices     

The design of wave digital filters using fully pipelined bit-level systolic arrays

Wave digital filters (WDF) based on analogue lattice and unit element cascade networks possess important properties that make them suitable for VLSI integration. These include low round-off noise even with short coefficient wordlength and a building block (two-port adaptor) with a very simple structure. Exploiting these properties leads to area efficient designs. Systolic architectures for these WDF networks give the additional advantage of very high sampling rates with potential application in sonar and video signal processing. The bit-level systolic array design of two WDFs is considered in detail, beginning at the filter specification and ending at the VLSI hardware architecture, with discussion of the expected values of the integrated circuit parameters.     

Phototransistor arrays of simplified design

Phototransistor arrays providing video output from a common collector electrode, with the emitters connected to the vertical address strips and the bases capacitively coupled to the horizontal strips, are shown to possess advantages in ease of fabrication and operation.     

Optimal virtual element patterns for adaptive arrays

A major problem in the use of small arrays is the perturbation in the element patterns due to mutual coupling and diffraction effects. The element patterns differ from each other and need to be corrected in order to obtain an array pattern that is close to the desired one. A matrix method can be used to correct the element patterns by modifying the original input/output weights to corresponding weights for corrected elements and vice versa. In this paper the main goal of the array correction is the maximal identity of the element patterns. In addition to identical element patterns the corrected array is characterized by uniform element spacing, which can be chosen to differ significantly from the element spacing in the real array making the corrected array more like a pure virtual array. For the array correction the linear least square error method has been used. To show the applicability of the virtual array, this method is applied to a typical beam scan case, and also over a frequency band.     

Optimal design of electrostatic lenses using integral equation field methods

The integral method of solving Laplace's or Poisson's equation results in a formulation for solution which is suitable for directly calculating the change in the domain potentials due to displacing the domain boundaries. Specifically, the order and locations remain the same for the matrix ∂Φ/∂x i.e., the change in potential per boundary change, which is required for the optimum design problem.     

Optimal matrix-filter design

A matrix filter produces N output values given a block of N input values. Matrix filters are particularly useful for filtering short data records (e.g. N⩽20). We introduce a new set of matrix-filter design criteria and show that the design of a matrix filter can be formulated as a convex optimization problem. Several examples are given of lowpass and bandpass designs as well as a Hilbert transformer design     

Ultrasonic imaging using arrays

An analysis is presented of the two general types of arrays-the direct imaging array and the phased array. These are analyzed over three levels of approximation: geometric optics, steady-state sinusoidal diffraction, and wide-band diffraction. Systems are studied which provide electronic deflection and focusing. In addition, systems are considered which exhibit high resolution in both lateral dimensions.     

Analytic design of conformal slot arrays

A completely analytic design process has been developed for small slot arrays which accounts for the varying effect of mutual coupling as a function of element position. A previously developed theory for the design of small arrays has been extended to include conformal dielectric-filled waveguide slot arrays. Computer software has been assembled which enables calculation of the slot geometries required to implement a specified aperture distribution and input impedance condition. The slot self- and mutual admittances are calculated numerically, thus eliminating the traditional measured slot data from the design process. This design technique has been applied to conformalX-band slot arrays on cylinders of a few wavelengths diameter. The arrays consist of multiple dielectric-filled waveguides, each of which is a narrow-band standing-wave linear array of longitudinal shunt slots. The computerized design process adjusts the length and offset of each slot in the total array until the desired aperture distribution and impedance match are achieved. A flow diagram of the design program and test results from experimental arrays are presented.     

Genetic design of linear antenna arrays

Discusses a way to optimize both the topology and the numerical parameters of an antenna design. The approach relies on an “antenna language” to define how antennas are constructed, and a genetic algorithm to create new designs using this language. The grammatical rules of a language can be very vague or very specific, depending on the purpose of the designer. With a vague grammar, genetic algorithms search a very large design space, and can occasionally find unexpected solutions to a design problem. Other times, they completely fail to find a reasonable solution because of the vastness of the search space. In this case, including knowledge about the problem into the grammar narrows the search to a region expected to yield good results. This yields more conventional design solutions that usually perform reasonably well. In an example, two languages were used to design a linear antenna array. The general language allowed a wide variety of designs, while the Yagi-log language confined the search to topologies known to perform well. The performance of the antennas produced by both languages was superior to that of a conventional log-periodic design. Further, the Yagi-log design was more fit than the unconventional design from the general language, illustrating the benefits of including knowledge in the grammar     

Corporate feed design for microstrip arrays

A design technique for an embedded microstrip corporate feed is presented. The aim of the design is to shape each corporate feed junction to achieve a tapered and in-phase output current distribution. From this, a Dolph-Chebyshev array sum pattern may be constructed. In the experiments, the designed corporate feed is used to excite a five-element equispaced linear array of overlay microstrip dipoles. It is found both theoretically and experimentally that a 20-dB Dolph-Chebyshev broadside sum pattern can be synthesized in the H-plane at 8.5 GHz. Mutual coupling among the dipoles is included in this analysis. The pattern degradation due to manufacturing tolerances on the alignment between the feed network and the dipole array is discussed     

The design of small slot arrays

The differences in mutual coupling for a central slot and a peripheral slot cannot be ignored in small arrays if good patterns and impedance are to be obtained. A theory has been developed whereby the length and offset of every slot in the array can be determined, in the presence of mutual coupling, for a specified aperture distribution and impedance match. The theory enlarges on Stevenson's method, and uses a modified form of Booker's relation based on Babinet's principle to treat nonresonant longitudinal shunt slots in the broad wall of a rectangular waveguide. A general relation between slot voltage and mode voltage is developed, and then formulas are derived for the active, self-, and mutual admittances among slots. These formulas result in a design procedure. Analogous treatments of inclined series slots in rectangular guide and of strip-line-fed slots are possible. Comparison between various experiments and the theory is presented. Tests of the theory include the resonant length of a zero offset slot, resonant conductance versus offset and resonant conductance versus frequency for a single slot, and self- and mutual admittances for two staggered slots. The design and performance of a two-by-four longitudinal shunt slot array is also described.     

Optimizing focused ion beam created solid immersion lenses in bulk silicon using design of experiments

In search of efficient solutions improving image resolution for backside failure analysis, the creation of solid immersion lenses in bulk silicon using focused ion beam has been investigated deeply in parametric detail. This technique is optimized using design of experiments, providing a better understanding of the pure ion beam sputtering process. It produces SILs in less than 20 min of processing time and offers an additional magnification of 3.2×. The optimal setup of this FIB created SIL is demonstrated and provides an improvement in resolution by a factor of 1.8. Limits of this technique are encountered and analyzed for future development.     

Optimal shape design of microwave device using FDTD and designsensitivity analysis

In this paper, a novel optimal shape design method is proposed using the finite-difference time-domain (FDTD) method and the design sensitivity analysis to obtain broad-band characteristics of microwave devices. In shape design problem, the nodes that describe the shape of geometry to be optimized are taken as design variables. The design sensitivity is evaluated using the adjoint variable equation that is obtained from a terminal-value problem. The adjoint equation can be solved by the FDTD technique with the backward time scheme. With this method, a Ka-band unilateral fin line is tested to show validity     

Optimal design of transform-based block digital filters using a quadratic criterion

Block digital filtering is a powerful tool to reduce the computational complexity of digital filtering systems. However, due to their block structure, block digital filters (BDFs) are time-varying linear systems, hence, their design is not easy. The most widely spread approaches to BDF design consist of constraining the BDF to be time-invariant (by restricting the design process to a specific subset of possible solutions) and then using conventional filter synthesis techniques. In this paper, we do not restrict the design process, and we propose a simple and optimal matrix-oriented approach to optimize the BDF coefficients. Furthermore, the proposed approach takes profit of the structure of transform-based BDFs to considerably reduce the computational complexity and memory requirements of the design process. Experimental results confirm that as expected, the obtained global distortion is lower than the distortion obtained with a traditional technique such as overlap-save.