A mathematical programming approach to a problem in variance penalised Markov decision processes

In this paper we present three algorithms for solving a problem in which it is required to get an optimal compromise between the average expected reward per unit time and the variance of the reward per unit time. The algorithms lead to an-optimal solution, where>0 is arbitrary.     

Dielectric properties of thin solid films formed on silicon

We report on the measurement of the frequency-dependent complex permittivity, ()=()-i(), over the frequency range, 30 MHz to 6 GHz, of silicon wafers and of thin dielectric films formed on silicon. Measurements, as a function of temperature and time treatments, were obtained by means of an HP Network Analyzer and dielectric probe and the resulting ()and()plots for the silicon wafers are shown to have a Debye-type [1] profile, thereby indicating that the associated polarization mechanism is of the orientational variety.     

Investigation of microstructure of molybdenum—copper black electrodeposited coatings with reference to solar selectivity

Optical and microstructural properties of electrodeposited molybdenum-copper (Mo-Cu) black coatings have been studied with reference to their selectivity in absorption of solar radiation. Such coatings were found to have a solar absorptance, , about 0.87 and low thermal emittance, , such that the selectivity, /, was 3.6. Electrodeposited molybdenum-black coatings generally have selectivity /3. The oxidation state of molybdenum in (Mo-Cu) black coatings as determined by X-ray photoelectron spectroscopy is about + 5 (which is fairly close to that of Mo₄O₁₁). Large numbers of irregular particles were found on the surface of molybdenum-copper black coatings. There is evidence that the particles contain copper oxide.     

The dielectric properties of aluminium nitride substrates for microelectronics packaging

The dielectric behaviour of sintered polycrystalline aluminium nitride substrates has been examined over the frequency range 500 Hz to 10 MHz and correlated with composition and microstructure. For pure, white AlN at 20 ° C both the permittivity () and dielectric loss () are frequency independent giving = 9.2±0.05 and tan = (2.1±0.1) × 10^–3. The permittivity is less than for pure alumina substrates ( = 10.2) but tan compares favourably, with that (1.4 × 10^–3) of alumina, which though used more widely has a thermal conductivity some eight times less than that of AlN. The addition of impurities, particularly iron, to give opaque black AlN causes large, frequency dependent increases in ; at 500 Hz the loss is seven times that of pure white AlN and is two times greater above 100 kHz. The temperature coefficient of permittivity [( – 1)( + 2)]^–1 [/T]_p between –180 and +180 ° C for pure white AlN is 1.05×10^–5 K^–1 which is similar to the value of 9×10^–6 K^–1 for pure Al₂O₃. For impure black AlN the coefficient below 20 ° C is the same but above 20 ° C there is a rapid, non-linear increase of with temperature. Below 180 ° C for pure white AlN and 20 ° C for impure black AlN the values of temperature coefficient are frequency independent at least up to 200 kHz.     

High-frequency dielectric behaviour of polycrystalline zinc substituted cobalt ferrites

The dielectric constant () and complex dielectric constant () of zinc substituted cobalt ferrites have been measured at room temperature in the high frequency range 100 kHz to 1 MHz. The values of dielectric loss tangent (tan ) have been computed from and . Plots of dielectric constant () versus frequency show a normal dielectric behaviour of the spinel ferrites. The frequency dependence of dielectric loss tangent (tan ) is found to be abnormal, giving a peak at certain frequency for all the ferrites under investigation. A qualitative explanation is given for the composition and frequency dependence of the dielectric constant and dielectric loss tangent. The dielectric constant for these mixed ferrites is approximately inversely proportional to the square root of the resistivity. A plot of dielectric constant versus temperature shows a transition near the Curie temperature. An attempt is made to explain the possible mechanism for this observation.     

Problem of measuring the thermophysical characteristics of thin resistive and dielectric films

We consider the theoretical and experimental possibilities of determining a complex of thermophysical characteristics of thin films on a substrate, on the basis of a nonstationary method with boundary conditions of the fourth kind.Notation time - x coordinate perpendicular to the film surface - ₂, ₀, ₁ coefficients of thermal activity (=/a) of the substrate, the film, and the secpnd body - T₀₀ initial temperature of the film-substrate system - T₁₀ initial temperature of the second body - =T₁₀ – T₀₀ T ₀ ⁰ temperature at the boundary after contact as 0 - T ₀ ^t8 temperature at the boundary after contact as Translated from Inzhenerno-Fizicheskii Zhurnal, Vol. 42, No. 1, pp. 98–101, January, 1982.     

Anomalous normal state and high temperature superconductivity in the cuprates

It is generally believed that a high electronic density-of-states n(E_F), and therefore a low Fermi velocity v_F, are required to obtain a high transition temperature, since T_{c ph} exp(–1/) and = n(E_F)V. However, V = < I² >/M _ph ² , and Bardeen showed that I E_Fk_F = (1/2)k _F ² v_F. Thus one may expect that should increase with v_F. While it may not be feasable to increase the one-electron velocity v_F signifcantly, the velocity may increase greatly as a result of renormalization by electron-electron interactions. Such a renormalization exists in Hartree-Fock theory for an unscreened electron-gas. We found that for a medium-density electron-gas (r_s 10–20) imbedded in a background with a dielectric constant () such that / > 10, there is a significant increase of v_F by renormalization, even when screening is taken into consideration. The peak of v(k) at k = k_F is very narrow, the half width being somewhat less than the frequency ₀ at which () falls by a factor of 2. When () is due to ionic polarization, _o is a typical phonon frequency. The height of the peak v_F/v _F ⁰ is of order E_F/₀, and the width in units of momentum is: k/k_F (₀/E_F)². This velocity peak is associated with a minimum in the screening constant at E_p. Its sharpness causes the normal state properties to be highly anomalous; namely the conductivity is exceptionally high, with an anomalous temperature dependence; the conductivity anisotropy, thermoelectric power, and Hall constant are anomalous. Direct determination of v(k) in YBCO by several methods indeed suggests a large, sharp peak, the width being of order 10 me V. This peak manifests itself in the tunneling and point-contact spectroscopy I–V curves. We suggest that the reduced screening at the Fermi level, associated with this velocity peak, is responsible for the high T_c of the cuprates and several other exotic superconductors. Thus, the high T_c is a reflection of the anomalous normal-state properties.     

Calculation of mixed-mode stress field at a sharp notch tip using <Emphasis Type="Italic">M</Emphasis><Subscript>1ɛ</Subscript>-integral

Direct computation of the mixed-mode stress field at a sharp notch tip appears to be difficult in that the mode I and mode II asymptotic stresses are in general governed by different orders of singularity. In this paper, we first present a path-independent integral termed M₁. The relation between M₁ and the generalized stress intensity factors is then derived and expressed as function of the notch angle. Once the M₁-integrals are accurately computed, the generalized SIF's and, consequently, the asymptotic mixed-mode stress field can thus be properly determined. No extra complementary solutions are required in the formulation. Further, no particular singular elements are required when the integration is performed by using finite elements. This work has been partially supported by National Science Council Grant No. NSC90-2211-E-008-040 to National Central University.     

Viscoelastic-plastic behaviour with mean strain changes in polypropylene

The stress-strain curves and stress-relaxation curves of polypropylene are obtained by using a closed-loop, electrohydraulic, servo-controlled testing machine. Effects of mean strain changes on deformation behaviour are examined in a tension-compression mode under strain control at room temperature (18–23 °C). The hysteresis loops of three mean strains show a steady-state response from the stress-strain curves at a strain rate of 1 × 10^–3 s^–1 at a strain width of 5%, at a number of cycles of N=50 and at three mean strains (_m=0, + 1.0 and + 2.0%). The drop of stress at the mean strain of _m= -1.0% is larger in magnitude than that at _m=+1.0%; this is caused by the higher stress level at _m=- 1.0% as compared with the stress level at _m=+1.0%. From the results of stress amplitude and the stress drop behaviour, the magnitude of stress drop is hardly affected by the mean strain.     

Sintering of strontium titanate in the presence of lithium salts in a reducing atmosphere

The possibility of preparing Type III ceramics for multilayer applications has been investigated in the strontium-titanium-oxygen system, using La₂O₃, Nd₂O₃and La_2/3TiO_3– as dopants and lithium salts as sintering agents. The introduction of bismuth was also explored. The sintering process is carried out by adding the lithium salts to mixtures of strontium titanate and dopants which are previously calcined in a reducing atmosphere. The dielectric properties and the microstructure of the ceramics have been studied. The best characteristics have been observed for the particular nominal composition SrTiO₃+0.75 mol% La_2/3Ti_3–+3 mol% Bi₂O₃+10 mol % LiNO₃leading to =20000, tan < 2%,R _i= 10¹¹ cm. These results are interpreted in terms of the formation of anionic vacancies and defects in the A sites of the perovskite ABO₃and of extended defects SrO and Bi₂O ₂ ²⁺ which are coherent with the perovskite matrix.     

Low endurance fatigue in metals and polymers

The fatigue behaviour of commercially pure aluminium and of nylon under sequentially varying strain amplitudes is compared with a damage law of the type suggested by Miner. Aluminium obeys such a law for both cyclic and uniaxial prestrains but the behaviour of nylon is significantly affected by microcracking, which produces a marked effect of loading sequence.Appendix N Number of strain cycles at a given time - N _f Value of N at failure - True tensile stress - True stress range for a strain cycled specimen - _h Value of at half the life of the specimen - True tensile strain - Total true strain range - _p True plastic strain range (= the breadth of the hysteresis loop at = 0) - _d True diametral strain range - E Young's modulus - Linear strain hardening rate when tested at a particular value of _p - D Damage due to cycling - D _p Damage due to prestrain - _p Prestrain. C, K, K₁, , are constants     

Environmental stress cracking of glassy polymers and solubility parameters

The use of solubility parameters to predict critical stress ( _c ^* ) or strains (_c) for environmental cracking/crazing in several glassy polymers (e.g. PMMA, PPO, PS, PVC, PSF and PC) is re-examined. It is shown that the enthalpic component ( _H ) of the Flory-Huggins semi-interaction parameter () does not always give a good correlation between _c and _H even though solvent molar volume and polymer-solvent molecular interactions have already been considered. Re-analysis of available experimental data using Gent's theory shows that there is a general trend for _c ^* (or _c) to increase with . These results, therefore, support Gent's proposed mechanism of environmental stress crazing/cracking. It is finally concluded that unless a definite relationship can be established between _c or _c ^* with _H it is not possible to predicta priori _c or _c ^* , given the empirical solubility parameters of a solvent. Unfortunately, there are not many such relationships as discovered in this paper.     

Direct and inverse heat-conduction problems for a semiinfinite rod for a partial outflow of heat through the surface

Analytical solutions of the direct and inverse problems of nonstationary heat conduction in a thin semiinfinite rod are given for the case of radiative heat fluxes at the lateral surfaces and a partial outflow of heat by convection and radiation through the end of the rod.Notation thermal diffusivity - x₁ coordinate along the length of the rod - t₁ time - t=t₁/d² dimensionless time (Fourier number) - x=X₁/d relative coordinate - T_o initial temperature - Boltzmann constant - Sk=aT_c ³d/ Stark number - Bi=d/ reduced Biot number - emissivity Translated from Inzhenero-Fizicheskii Zhurnal, Vol. 47, No. 1, pp. 148–153, July, 1984.     

Dynamic scaling and ultrasonic attenuation in 4He near the superfluid transition point

Attenuation of first sound has been measured in ⁴He under saturated vapor pressure near the lambda temperature T at frequencies /2 ranging from 10.2 to 271 MHz. The frequency dependence of the critical part of the attenuation is determined and the dynamic scaling hypothesis is examined. Above the lambda point, it is found that the critical attenuation is described by a scaling function (, ) = ^1+y F(), where = ₀^–x and = T/T – 1, with the results x = 1.02±0.05 and y = 0.33±0.03. The characteristic frequency of the order-parameter fluctuation with the wave number k equal to the inverse correlation length is then proportional to ^x , which is in an excellent agreement with the prediction of dynamic scaling. Below the lambda point, a characteristic relaxation time or times shorter than previously expected at lower frequencies appears to exist in the present frequency range.Based on a Ph.D dissertation submitted by K. Tozaki to the University of Tokyo (1977).     

Effective Łojasiewicz inequalities in semialgebraic geometry

The main result of this paper can be stated as follows: letV ⁿ be a compact semialgebraic set given by a boolean combination of inequalities involving only polynomials whose number and degrees are bounded by someD > 1. LetF, G[X₁,, X_n] be polynomials with degF, degG D inducing onV continuous semialgebraic functionsf, g:VR. Assume that the zeros off are contained in the zeros ofg. Then the following effective ojasiewicz inequality is true: there exists an universal constantc ₁ and a positive constantc ₂ (depending onV, f,g) such that for allxV. This result is generalized to arbitrary given compact semialgebraic setsV and arbitrary continuous functionsf,g:V . An effective global ojasiewicz inequality on the minimal distance of solutions of polynomial inequalities systems and an effective Finiteness Theorem (with admissible complexity bounds) for open and closed semialgebraic sets are derived.     

Thermal transport properties in helium near the superfluid transition. I.4He in the normal phase

The thermal conductivity and the associated relaxation time to reach steady-state conditions are reported for the normal phase of several very dilute mixtures of³He in⁴He (X<4 × 10^–6) at saturated vapor pressure near T. The measurements were made over the reduced temperature range 2.5 × 10^–6<<2×10^–1, where (T–T)/T, and are representative for pure⁴He. The spacing between the cell plates was 0.147 cm. The systematic uncertainty in the conductivity data is estimated to increase from 2% for =0.2 to 4% for =3 × 10^–6. The random scatter due to finite temperature resolution increases to 7% at the smallest . The data are in agreement within the combined uncertainty with recent ones by Tam and Ahlers (cell F, spacing 0.20 cm) and with previous ones in this laboratory taken with a different plate spacing. The thermal diffusivity coefficientD _T = / C _p obtained from is found to agree within better than 15% with the calculated one using data for , the density , and the specific heatC _p . Measurements of the effective boundary resistivityR _b in the superfluid phase are described.R _b is found to depend on the thermal history of the cell when cycled up to 77 K and above. Also,R _b shows the beginning of an anomalous increase for ¦¦10^–4. The possible reasons for this anomaly are discussed, and their impact on the analysis of conductivity data in the normal phase is appraised.     

Effective permittivity of layered dielectric sphere composites

A straightforward model is presented for analysing the effective permittivities of layered dielectric sphere composites. Using the present model, the effective permittivity, _eff, of layered dielectric sphere composites can be deduced using classical two-phase dielectric mixture formulae in two steps: first, the effective permittivity, _incl, of the inclusions is calculated by taking the layered dielectric sphere inclusions as sub-composites; and second, the effective permittivity, eff, of the composites is found by substituting the layered dielectric sphere inclusions with homogeneous spheres whose permittivity is equal to _incl. The present model is applicable to multi-layer sphere composites. Experiments on resin-based hollow bead composites show that the present model accurately predicts the effective permittivity of layered dielectric sphere composites.     

Longitudinal dielectric function and collective surface oscillations in superconductors

The electrodynamic approach to the problem of collective oscillations in superconductors is considered. The kinetic equation for quasiparticles is used to find the longitudinal dielectric function ₁ (, k) of a pure superconductor at ( is the collision frequency). With the help of the evaluated ₁ (, k) the existence of collective oscillations localized near the boundary of the superconductor is investigated.     

An analysis of the conditions of brittle fracture origin

A modified criterion of brittle fracture, a necessary condition of which, the condition of origin of microcracks, is plastic deformation until some value of ₀ dependent upon temperature, is experimentally substantiated. An experimental-calculation method is developed for determination of the relationship of ₀ to temperature T, on the basis of which the ₀(T) relationship is obtained for type 15Kh2MFA steel. It is shown that the proposed formulation of the criterion of brittle fracture makes it possible to consistently describe the fracture processes with different stressed and strained states of the material.Leningrad Branch, Institute of Mechanical Engineering. Translated from Problemy Prochnosti, No. 11, pp. 9–13, November, 1989.     

Measurement of temperature in a non-steady-state gas flow

A method is described for measuring the temperature of a non-steady-state gas flow with a thermocouple which is an inertial component of the first order.Notation T*_f non-steady-state gas flow temperature - T_t thermosensor temperature - thermal inertia factor of thermosensor - time - C total heat capacity of thermosensor sensitive element - S total heat-exchange surface between sensitive element and flow - heat-liberation coefficient - temperature distribution nonuniformity coefficient in sensitive element - Re, Nu, Pr, Bi, Pd hydromechanical and thermophysical similarity numbers - P^* total flow pressure - P static flow pressure - T^* total flow temperature - d_t sensitive element diameter - w gas flow velocity - flow density - flow viscosity - _f flow thermal conductivity - k gas adiabatic constant - R universal gas constant - M Mach number - T thermodynamic flow temperature - _o, _o and values at T=288°K - A, m, n, p, r coefficients - _c heat-liberation coefficient due to colvection - _r heat-liberation coefficient due to radiation - _b emissivity of sensitive element material - Stefan-Boltzmann constant - T_e temperature of walls of environment - _c, _r, _tc thermosensor thermal inertia factors due to convective, radiant, and conductive heat exchange - L length of sensitive element within flow - a thermal diffusivity of sensitive element material - _t thermal conductivity of sensitive element material Translated from Inzhenerno-Fizicheskii Zhurnal, Vol. 47, No. 1, pp. 59–64, July, 1984.