Characterization and Control of Residual Stress and Curvature in Anodically Bonded Devices and Substrates with Etched Features

While anodic bonding is commonly used in a variety of microelectromechanical systems (MEMS) applications, devices and substrates that incorporate this processing technique are often subjected to significant residual stress and curvature that create post-processing and reliability issues. Here, using an anisothermal anodic bonding procedure, residual stresses and the resulting wafer curvature in these structures are controlled by varying the initial bond temperatures of the silicon and Pyrex wafers independently. Residual stresses are quantified by measuring bulk wafer curvature and, locally, stress concentrations are measured using infrared photoelasticity accompanied by 3-D thermomechanical finite element analysis. Based on the good agreement between numerical predictions and experimental results, this process can be used to determine the bulk post-bond wafer curvature and to reduce the likelihood of structural failure at these sites, by changing the residual stresses from tensile in nature, which may drive initiation and growth of cracks, to compressive, which can suppress such failures.     

Rectangular convexity of convex domains of constant width

An E R ² is r-convex if for every x, y E there exists a closed rectangle R such that x, y R and R E. Several results about r-convexity appeared in [1]. Its authors formulated a conjecture about conditions for a compact, convex set in R ² to be r-convex. We prove this conjecture in the case of convex domains of constant width.     

Facets with fewest vertices

Forv>d≧3, letm(v, d) be the smallest numberm, such that every convexd-polytope withv vertices has a facet with at mostm vertices. In this paper, bounds form(v, d) are found; in particular, for fixedd≧3, $$\frac{{r - 1}}{r} \leqslant \mathop {\lim \inf }\limits_{\upsilon \to \infty } \frac{{m(\upsilon ,d)}}{\upsilon } \leqslant \mathop {\lim \sup }\limits_{\upsilon \to \infty } \frac{{m(\upsilon ,d)}}{\upsilon } \leqslant \frac{{d - 3}}{{d - 2}}$$ , wherer=[1/3(d+1)].     

Hypersurfaces of the Euclidean sphere with nonnegative Ricci curvature

In this paper we prove that a compact oriented hypersurface of a Euclidean sphere with nonnegative Ricci curvature and infinite fundamental group is isometric to an H(r)-torus with constant mean curvature. Furthermore, we generalize, whithout any hypothesis about the mean curvature, a characterization of Clifford torus due to Hasanis and Vlachos. Received: 19 March 2002     

An Approach to Fuzzy Noncooperative Nash Games

Systems that involve more than one decision maker are often optimized using the theory of games. In the traditional game theory, it is assumed that each player has a well-defined quantitative utility function over a set of the player decision space. Each player attempts to maximize/minimize his/her own expected utility and each is assumed to know the extensive game in full. At present, it cannot be claimed that the first assumption has been shown to be true in a wide variety of situations involving complex problems in economics, engineering, social and political sciences due to the difficulty inherent in defining an adequate utility function for each player in these types of problems. On the other hand, in many of such complex problems, each player has a heuristic knowledge of the desires of the other players and a heuristic knowledge of the control choices that they will make in order to meet their ends.In this paper, we utilize fuzzy set theory in order to incorporate the players' heuristic knowledge of decision making into the framework of conventional game theory or ordinal game theory. We define a new approach to N-person static fuzzy noncooperative games and develop a solution concept such as Nash for these types of games. We show that this general formulation of fuzzy noncooperative games can be applied to solve multidecision-making problems where no objective function is specified. The computational procedure is illustrated via application to a multiagent optimization problem dealing with the design and operation of future military operations.     

Intra- and intersubject variability in acoustic measures of normal voice

Twenty-four normal adult women read part of the Rainbow Passage and sustained vowels three trials each. Utterances were assessed for selected parameters measured by Visi-Pitch (average and SD of fundamental frequency (F₀), average and SD of dBA, perturbation, and percent voiced/unvoiced/pause). Assessment of each parameter included measures of central tendency, dispersion, and distribution characteristics (skewness and kurtosis) of the data and of the ranges of values that would include 95% of the scores (95% fiduciary limits). Generally, differences for the group between the three trials were not significant. Intersubject variability for only a few parameters was less than 20% of the parameter's mean. For vowels, variability of jitter was 30–48% of the mean. Eight subjects provided performances 2 months later to obtain an estimate of intrasubject variability over time. There were desirable intrasubject correlations between performances for mean F₀, jitter in reading and on vowels /i/ and /a/, and percent of voicing. Inter- and intrasubject variability seems restricted and the data appear to resemble a normally distributed function for mean F₀ on reading, jitter on /i/, and percent of voicing. Thus, these parameters may have statistical merit for use in vocal testing.