摘要:

AbstractThe Runge‐Kutta method, a popular numerical procedure for solving a differential equation, is based on Taylor series expansions of the solution. It provides the solution if the function is single‐valued. However, it is difficult to obtain the solution in the case of a multivalued function. In this paper, a coordinate system that follows the shape of the solution is introduced for developing a numerical solution of a differential equation with a multivalued functional solution. For a step describing accurately the shape of the function, the arc length of the solution function is used. the Runge‐Kutta method with arcs of equal length is proposed in which the arc length is kept identical. From the error formula for the Runge‐Kutta method, the condition is shown under which the solution by the proposed method is more accurate than the one by the conventional Runge‐Kutta method. Next, a numerical method for a differential equation with a multivalued solution function is given by combining the new Runge‐Kutta method and the coordinate transformation. This transformation is applied not only to numerical values but also to the differential equation itself. This process can eliminate the points at which the function on the right‐hand side of the differential equation becomes singular and provide solutions even when the solution function i

ISSN:8756-6621    

DOI:10.1002/ecja.4400660302

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractThis paper considers the necessary and sufficient condition for the state variables for locally solvable nonlinear networks to be capacitor currents vectoriCand inductor voltagesvL, (iCT, vLT), instead of capacitor charges vectorqand inductor fluxes vector ψ (qT, ψT). the condition is given by the regularity of matrixKand |ta, vj|→∞ as |e, j|→∞ in the resistance network where all the capacitors and inductors are replaced by voltage sourceseand current sourcesj, respectively. Here,ieandvjare the voltage source current and current source voltage, respectively.As a result, in the networks where (qt, ψt) can be transformed into (iCT, vLT), the network dynamics can be described in the same way as in a linear system and global consideration on networks can be done only by investigating the property of matrixKA.For example, the sufficient condition for the equilibrium point to be globally asymptotic‐stable is given comparatively easily, the consideration on eventual passivity can be done easily, and, in general, there exist various analysis methods such as in the case where (qT, ψT) are used as st

ISSN:8756-6621    

DOI:10.1002/ecja.4400660303

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractStudies on the analysis ofM/G/1 andGI/M/1 using the supplementary variable approach have already been reported. However, there has been no study which analyzed consistently the systems from the transient state to the steady state.Using the remaining service time of a customer being served in theM/G/1 and the residual interarrival time in theGI/M/1 as supplementary variables, we derive the basic equations of state probability density functions and state transition probability density functions which describe the behavior of the systems and the transient distributions of characteristic quantities, such as queue length, virtual waiting time, etc. the transient solution is derived as a solution of certain initial‐value‐ and boundary‐value problems. Using this solution, the transient solutions of the characteristic quantities (including queue length, virtual waiting time, mean queue length, mean virtual waiting time, etc.) are derived. Furthermore, by applying the final value theorem of Laplace transform to these results, we obtain stationary distributions of the characteristic quantities which equal those results obtained by imbedded Markov chain ana

ISSN:8756-6621    

DOI:10.1002/ecja.4400660304

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractIn some cases of realizing a circuit on a single‐layer printed board, all modules are mounted on the same side of the board. This paper proposes an algorithm, assuming that a circuit contains modules which cannot be mounted upside‐down, to determine whether or not a single‐layer printed wiring is possible under the following two restrictions: (i) the modules can be mounted only on the specified side of the board; (ii) wiring should not be made under the module. the notion ofPQtree is extended as the data structure in the proposed algorithm, and the computational time is proportional to the total number of terminals in the ci

ISSN:8756-6621    

DOI:10.1002/ecja.4400660305

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractThe authors previously described the design of shielded TM010dielectric rod resonators (below referred to as “TM010resonators”), a new configuration of small‐size resonators for microwave frequencies. This paper describes the design of bandpass filters using coaxially oriented electrically‐coupled TM010dielectric rod resonators. First, the resonant frequency of electrically‐coupled TM010resonators is analyzed by applying the Ritz‐Galerkin method to the integral equations that express the boundary conditions. This method is shown to be effective in numerical calculation of the coupling coefficients and center frequency of the coupled resonators and the resonant frequencies for inserted TM010resonators which are required for filter design. Next the characteristics of the coupling probe which effectively produces the strong coupling of the axially symmetric TM010mode of resonators and suppresses the excitation of the higher modes are determined experimentally, making it possible to estimate the center frequency for a filter which takes the variation of the resonant frequency produced by the excitation and coupling hole into consideration. Finally, design of 2‐ and 3‐stage filters of various bandwidths at 2.7 GHz demonstrates the suitability of the approach. Relative bandwidths of 1 to 10%, low losses of. 15 to. 6 db, and absence of spurious responses over bands wider than 1 octa

ISSN:8756-6621    

DOI:10.1002/ecja.4400660306

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractThe amount of network control as well as the traffic efficiency should be considered in designing a network satisfying the overall economy of the switching system. If high‐speed PCM is adopted,T‐S‐Tnetworks are preferable from the traffic efficiency viewpoint. However,S‐T‐Snetworks can be preferable from the viewpoint of the amount of control. In this paper, a preferable form ofS‐T‐Stime division networks is derived from the results of the simulation of the network control processes.S‐T‐Stype networks are compared withT‐S‐Ttype networks from the viewpoints of both the traffic efficiency and the amount of control, and it is revealed thatS‐T‐Stype networks require less control. Finally, a desired network form satisfying the overall economy is clarified in terms of the whole time division networks including bot

ISSN:8756-6621    

DOI:10.1002/ecja.4400660307

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractIt is very important from the viewpoint of environmental electromagnetic engineering to detect vectors of high‐frequency currents on a metal plate (high‐frequency surface currents). In this paper, a high‐frequency surface current detector has been realized based on the coupling theory between an electromagnetic plane wave and a wire above the earth, and by modelling this arrangement. For the modelling, it has been considered that the electric field is approximately perpendicular to the metal plane in the space near its surface due to the surface currents flowing along the plate; and that the magnetic field is parallel to the metal plane. In practice, the pointing vector carried by the magnetic field is measured. It has been confirmed that the detector actually made can detect the vectors of surface currents by detecting progressing waves generated along a metal plane. the coupling coefficient between the surface currents and detector has been obtained experimentally by varying the parameters of the coupling

ISSN:8756-6621    

DOI:10.1002/ecja.4400660308

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractThis paper proposes new structures for a polarization‐maintaining optical fiber with axially asymmetric index distribution: (1) a single circular‐pit fiber (SCF), (2) a double circular‐pit fiber (DCF), (3) a non‐coaxial W‐type fiber (NWF). Numerical analyses of their propagation characteristics are based on a new point‐matching method whose computation error is limited only by the mode number truncation error. It is shown that values of β/k0as large as 3.0 × 10−4can be obtained with air pits in the SCF and

ISSN:8756-6621    

DOI:10.1002/ecja.4400660309

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractA theoretical analysis based on the multiple scale method is presented for the behavior of the leaky wave in a corrugated ferrite image guide. the attenuation of the guided mode due to leakage is given analytically as a complex perturbation of the propagation constant. As an application of this structure, a millimeter wave leaky wave antenna has been considered and its radiation efficiency and directivity have been studied. Next, experiments are carried out for a leaky wave antenna using a poly‐crystalline YIG slab. the radiation pattern has been measured in the 40‐GHz band and compared with theoretical values. A change in dc magnetic bias by 1 Wb/m2shifts the main beam to be scanned by 17.5 degrees. the experimental results agree well with theoretical d

ISSN:8756-6621    

DOI:10.1002/ecja.4400660310

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY

摘要:

AbstractWhen a Howells‐Applebaum adaptive array is used in a communication system, the receiving side must know the frequency and arrival angle of the desired signal. However, it is often impractical to assume that the arrival angle is accurately known. It is of importance to determine the degradation of output characteristics due to the angular error in arrival angle of the desired signal. This paper first presents a theoretical analysis of the steady‐state performance of the array when only the desired signal exists and angular error is contained in it. A condition is found under which the desired SNR is obtainable from the angle error profile. A formula is derived by which the upper limit of the amplifier gain or the allowable upper limit in the angle error is obtained. These limits are important criteria in the array design. Some numerical calculations are conducted for angle error characteristics in the presence of interference and the effectiveness of the derived design criteria. It is found that these criteria are sufficiently useful if the amplifier gain g normalized by the thermal noise power is larger than unity or if for a small g the interference power is reasonably large. This is because the angle error characteristics are better than those in the absence of interfere

ISSN:8756-6621    

DOI:10.1002/ecja.4400660311

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1983

数据来源: WILEY