This paper presents a new connected associative memory neural network. In this network, a threshold function which has two dynamical parameters is introduced. After analyzing the dynamical behaviors and giving an upper bound of the memory capacity of the conventional connected associative memory neural network, it is demonstrated that these parameters play an important role in the recalling processes of the connected neural network. An approximate method of evaluationg their optimum values is given. Further, the optimum feedback stopping time of this network is discussed. Therefore, in our network, the recalling processes are ended at the optimum feedback stopping time whether a state energy has been local minimum or not. The simulations on computer show that the dynamical behaviors of our network are greatly improved. Even though the number of learned patterns is so large as the number of neurons, the statistical properties of the dynamical behaviors of our network are that the output series of recalling processes approach to the expected patterns on their initial inputs.

We propose Polarization-Shift-Keying (POLSK) homodyne system using phase-diversity receivers and theoretically analyze its bit-error-rate (BER) performance. Since the proposed system uses polarization modulation and homodyne detection, it can cancel the phase noise and is attractive at a high bit-rate transmission. It is found that the receiver sensitivity of the proposed POLSK homodyne system is the same as that of POLSK heterodyne system and is much better than that of DPSK phase-diversity homodyne systems at high signal-to-noise ratio (SNR). We also cosider theoretically the effect of the fluctuation of state of polarization (SOP) on the BER performance of POLSK homodyne system.

We consider applications of optical solitons to signal processing. Soliton switching devices promise ultrafast operation and compatibility with communications systems using optical pulses. Quantum soliton effects include broadband squeezing and quantum nondemolition measurements, and can reduce noise and increase sensitivities of optical measurements. We report the demonstration of two-color soliton switching and describe progress towards implementation of quantum nondemolition measurement of photon number using soliton collisions.

We consider applications of optical solitons to signal processing. Soliton switching devices promise ultrafast operation and compatibility with communications systems using optical pulses. Quantum soliton effects include broadband squeezing and quantum nondemolition measurements, and can reduce noise and increase sensitivities of optical measurements. We report the demonstration of two-color soliton switching and describe progress towards implementation of quantum nondemolition measurement of photon number using soliton collisions.

We study the depth of planar Boolean circuits. We show that planar Boolean circuits of depth D(n) are simulated by on-line Turing machines in space O(D(n)). From this relationship, it is shown that any planar circuit for computing integer multiplication requires linear depth. It is also shown that a planar analogue to the NC-hierarchy is properly separated.

The nonlinear optical properties of organics with unsaturated bonds were compared with those of inorganics including semiconductors and dielectrics. Because of the mesomeric effect, namely quantum mechanical resonance effect among configurations, aromatic molecules and polymers have larger optical nonlinear parameters defined as δ(n)X(n)/(X(1))n both for the second (n2) and third-order (n3) nonlinearities. Experimental results of ultrafast nonlinear response of conjugated polymers, especially polydiacetylenes, were described and a model is proposed to explain the relaxation processes of photoexcitations in the conjugated polymers. Applying the model constructed on the basis of the extensive experimental study, we propose model polymers to obtain ultrafast resonant optical nonlinearity.

Several issues on semiconductor lasers with low dimensional quantum systems are discussed. First, described are fabrication techniques for quantum wire and box structures, particularly a selective growth MOCVD growth technique which have been recently developed. Using this technique, we obtained 20 nm15 nm triangular-shaped quantum wire structures. Next, we investigate band structures of the quantum wires having strain effects, including lasing characteristics of quantum wire lasers with the strain effects. Finally we discuss importance to control both the electron wave mode and the optical wave mode for future high performance lasers, which leads to the concept of quantum micro-lasers. In order to demonstrate possibility to control the spontaneous mode in the laser cavity, an experimental result is shown on enhancement and inhibition effects of the spontaneous emission mode in a vertical cavity laser having two kinds of the quantum well.

In this paper we study the problem of scheduling a tree-structured program on multiprocessors so as to minimize the total execution time, which includes communication delay between processors. It is assumed in the problem that a sufficiently large number of processors are available. It is known that if the program structures are restricted to be out-trees, the problem can be solved in O(n2) time, where n denotes the number of modules of a program. However, this problem is known to be NP-hard if the program structures are allowed to be in-trees. Up to now, no optimal algorithm, except an obvious one, was known for the latter case while some approximation algorithms were shown. We present an optimization algorithm with a nontrivial time bound O((1.52)nn log n) for the in-tree case.

We propose Polarization-Shift-Keying (POLSK) homodyne system using phase-diversity receivers and theoretically analyze its bit-error-rate (BER) performance. Since the proposed system uses polarization modulation and homodyne detection, it can cancel the phase noise and is attractive at a high bit-rate transmission. It is found that the receiver sensitivity of the proposed POLSK homodyne system is the same as that of POLSK heterodyne system and is much better than that of DPSK phase-diversity homodyne systems at high signal-to-noise ratio (SNR). We also cosider theoreically the effect of the fluctuation of state of polarization (SOP) on the BER performance of POLSK homodyne system.

Transmittance distribution along a horizontal line in LCDs addressed by amorphous silicon TFTs was investigated using measurements and calculations. Nonuniformity of the distribution, in which the transmittance increased with increasing distance from the left edge of the LCD, was observed in a 10 inch diagonal TFT-LCD. The cause of the nonuniformity was attributed to the decrease in voltage drop due to the gate source parasitic capacitance and the increase in gate voltage fall time due to large line resistance, based on the measurements of voltage drops in TFT test elements and calculations considering the decrease in voltage drop. The distribution could be improved by reducing the line resistance and parasitic capacitance in the actual LCD.

Suppose that there are terminals on two concentric circles Cin and Cout, with Cin inside of Cout. A set of two-terminal nets is given and the routing area is the annular region between the two circles. In this paper, we present an O(n2) time algorithm for testing whether the given net set is two-layer routable, where n is the number of nets. Applying this algorithm repeatedly, we can find, in O(n3) time, a maximal subset of nets which is two-layer routable.

This paper presents an approximation method for deriving the availability of a parallel redundant system with preventive maintenance (PM) and common-cause failures. The system discussed is composed of two identical units. A single service facility is available for PM and repair. The repair times, the PM times and the failure times except for common-cause failures are all assumed to be arbitrarily distributed. The presented method formulates the problem of the availability analysis of a parallel redundant system as a Markov renewal process which represents the state transitions of one specified unit in the system. This method derives the availability easily and accurately. Further, the availability obtained by this method is exact in a special case.

The VSTEP concept and its practical application in the form of an LED-type pnpn-VSTEP demonstrating low power consumption through electro-photonic operational modes are both shown. Further, with focus primarily on the new laser-mode VSTEP with high-intensity light output and narrow optical beam divergence, the design features such as threshold gain and optical absorptivity, device fabrication, and characteristics are explained. The possibility of ultimate performance based mainly on electrical to optical power conversion efficiency, important from the application viewpoint of optical interconnection, are also discussed. Also, as two examples of functional optical interconnection achieved by VSTEP, serial-to-parallel data conversion and optical self-routing switches are shown. Finally, future opto-electronic technologies to be developed for two-dimensionally integrable surface-type optical semiconductor devices, including the VSTEP, are discussed.

In this paper, some classes of arithmetic circuits are introduced that capture the computational complexity of computing the determinant of matrices with entries either indeterminates or constants from a field. An arithmetic circuit is just like a Boolean circuit, except that all AND and OR gates (with fan-in two) are replaced by gates realizing a multiplication and an addition, respectively, of two polynomials over some indeterminates with coefficients from the field, and the circuit computes a (formal multivariate) polynomial in the obvious sense. An arithmetic circuit is said to be skew if at least one of the inputs of each multiplication gate is either an indeterminate or a constant. Then it is shown that for all square matrices M of dimension q, the determinant of M can be computed by a skew arithmetic circuit of (q20) gates, and is shown that for all skew arithmetic circuits C of size q, the polynomial computed by C can be defined as the determinant of a square matrix M of dimension (q). Thus the size of skew arithmetic circuit is polynomially related to the dimension of square matrices when it is considered to represent multivariate polynomials in both arithmetic circuits and the determinant. The results are extended to some other classes of arithmetic circuits less restricted than skew ones, and by using such an extended result, a difference and a similarity are demonstrated between polynomials represented as the determinant of matrix of relatively small dimension and those polynomials computed by arithmetic formulas and arithmetic circuits of relatively small size and degree.

In this paper we review the recent progress and basic technology of vertical cavity surface emitting lasers together with related parallel surface operating optical devices. First, the concept of surface emitting lasers is presented, and then currently developed device technologies will be reviewed. We will feature several technical issues, such as multi-layer structures, 2-dimensional arrays, photonic integration, etc. Lastly, future prospects for parallel lightwave systems will be discussed.

A chordal ring network is a processor network on which n processors are arranged to a ring with additional chords. We study a distributed leader election algorithm on chordal ring networks and present trade-offs between the message complexity and the number of chords at each processor and between the message complexity and the length of chords as follows:For every d(1dlog* n1) there exists a chordal ring network with d chords at each processor on which the message complexity for leader election is O(n(log(d1)nlog* n)).For every d(1dlog* n1) there exists a chordal ring network with log(d1)nd1 chords at each processor on which the message complexity for leader election is O(dn).For every m(2mn/2) there exists a chordal ring network whose chords have at most length m such that the message complexity for leader election is O((n/m)log n).

A reversible (or injective) cellular automaton (RCA) is a backward deterministic" CA, i.e., every configuration of it has at most one predecessor. Margolus has been shown that there is a computation-universal two-dimensional 2-state RCA model. Although his model is very interesting, it differs from a standard CA model because of its somewhat spatial and temporal non-uniformity. In this paper, we present two kinds of simple 16-state computation-universal models using the framework of two-dimensional reversible partitioned CA (PCA). Since PCA can be considered as a subclass of standard CA, we can immediately obtain 16-state standard RCA models from them. For each of these models, we designed a configuration which simulates a Fredkin gate. Since Fredkin gate has been known to be a universal logic element, computation-universality of these two models is concluded.

The VSTEP concept and its practical application in the form of an LED-type pnpn-VSTEP demonstrating low power consumption through electro-photonic operational modes are both shown. Further, with focus primarily on the new laser-mode VSTEP with high-intensity light output and narrow optical beam divergence, the design features such as threshold gain and optical absorptivity, device fabrication, and characteristics are explained. The possibility of ultimate performance based mainly on electrical to optical power conversion efficiency, important from the application viewpoint of optical interconnection, are also discussed. Also, as two examples of functional optical interconnection achieved by VSTEP, serial-to-parallel data conversion and optical self-routing switches are shown. Finally, future opto-electronic technologies to be developed for two-dimensionally integrable surface-type optical semiconductor devices, including the VSTEP, are discussed.

This paper investigates the performance of the line spectrum pair (LSP) frequency parameter representation for speech recognition. Transitional parameters of LSP frequencies are defined using first-order regression coefficients. The transitional and the instantaneous frequency parameters are linearly combined to generate a single feature vector used for recognition. The performance of the single vector is compared with that of the cepstral coefficients (CC) representation using a minimumdistance classifier in speaker-independent isolated word recognition experiments. In the speech recognition experiments, the transitional and the instantaneous coefficients are also combined in the distance domain. Also, inverse variance weighted Euclidean measures are defined using LSP frequencies to achieve Mel-scale-like warping and the new warped-frequencies are used in recognition experiments. The performance of the single feature vector defined with transitional and instantaneous LSP frequencies is found to be the best among the measures used in the experiments.

In this paper we review the recent progress and basic technology of vertical cavity surface emitting lasers together with related parallel surface operating optical devices. First, the concept of surface emitting lasers is presented, and then currently developed device technologies will be reviewed. We will feature several technical issues, such as multi-layer structures, 2-dimensional arrays, photonic integration, etc. Lastly, future prospects for parallel lightwave systems will be discussed.