The semantics of a language for communicating processes is investigated, and three full abstractness results for are established. The language contains atomic actions, termination, inaction, sequential composition, alternative composition, parallel composition, action restriction, and a form of guarded recursion. (The guardedness restriction on recursion is needed to establish one of the full abstractness results.) Three Plotkin-style operational semantics WL, CWL, and IWL of the language are defined. These semantics are linear in that the meaning of each program in any of these semantics is a set of action sequences the program may perform, and are weak in that the action sequences are obtained by ignoring (finte sequences of) internal moves. All the three semantics distinguish divergence (an infinite sequence of internal moves) from deadlock. The semantics IWL differs from the other two in that IWL is a so-called interal action semantics taking into account only internal moves under the assumption that the environment allows no (external) communication actions, and hence, the only possible actions for processes are internal moves, whereas the other two semantics take into account communication actions in addition to internal moves. The two semantics WL and CWL differ each other in that CWL is a so- called completed trace semantics, whereas WL is not. Then, two compositional models RF and IRF for the language are proposed, and the full a bstractness of RF (resp. of IRF) w. r. t. WL and CWL (resp. w. r. t. IWL), as expressed in the following, is established:s1s2(*) S[][S[]is a context of ⇒S[s1]S[s2]],where ,RF,WL, RF,CWL, IRF,IWL. A similar full abstractness result has been established by Bergstra, Klop, and Olderog for a language without recursion or internal moves. Moreover, Rutten investigated the semantics of a language similar to , in the framework of complete metric spaces, and showed that the failures model is fully abstract w. r. t. a strong linear semantics L, where L is strong in that it does not abstract from internal moves. The full abstractness of RF w. r. t. CWL, expressed in (*) with ,RF,CWL, is an extension of the result of Bergstra et al. to a language with recursion and internal moves. Also, the full abstractness of IRF w. r. t. IWL, expressed in (*) with ,IRF,IWL, is an extension of the Rutten's result, to the case of weak linear semantics with divergence.

By adding some functions to memories, highly parallel computation may be realized. We have proposed memory-based parallel computation models, which uses a new functional memory as a SIMD type parallel computation engine. In this paper, we consider models with communication between the words of the functional memory. The memory-based parallel computation model consists of a random access machine and a functional memory. On the functional memory, it is possible to access multiple words in parallel according to the partial match with their memory addresses. The cube-FRAM model, which we propose in this paper, has a hypercube network on the functional memory. We prove that PSPACE is accelerated to polynomial time on the model. We think that the operations on each word of the functional memory are, in a sense, the essential ones for SIMD type parallel computation to realize the computational power.

It is known that the problem of finding a largest common subgraph is NP-hard for general graphs even if the number of input graphs is two. It is also known that the problem can be solved in polynomial time if the input is restricted to two trees. In this paper, a randomized parallel (an RNC) algorithm for finding a largest common subtree of two trees is presented. The dynamic tree contraction technique and the RNC minimum weight perfect matching algorithm are used to obtain the RNC algorithm. Moreover, an efficient NC algorithm is presented in the case where input trees are of bounded vertex degree. It works in O(log(n1)log(n2)) time using O(n1n2) processors on a CREW PRAM, where n1 and n2 denote the numbers of vertices of input trees. It is also proved that the problem is NP-hard if the number of input trees is more than two. The three dimensional matching problem, a well known NP-complete problem, is reduced to the problem of finding a largest common subtree of three trees.

There have been several studies related to a reduction of the amount of computational resources used by Turing machines. As consequences, Linear speed-up theorem", tape compression theorem" and reversal reduction theorem" have been obtained. In this paper, we discuss a leaf reduction theorem on alternating Turing machines. Recently, the result that one can reduce the number of leaves by a constant factor without increasing the space complexity was shown for space- and leaf-bounded alternating Turing machines. We show that for time- and leaf-bounded alternating Turing machines, the number of leaves can be reduced by a constant factor without increasing time used by the machine. Therefore, our result says that a constant factor on the leaf complexity does not affect the power of time- and leaf-bounded alternating Turing machines.

The heterointerfaces of Al0.3Ga0.7As/GaAs single quantum wells (SQWs) and the characteristics of SQW lasers grown on Si substrates with Al0.5Ga0.5As/Al0.55Ga0.45P intermediate layers (AlGaAs/AlGaP ILs) entirely by metalorganic chemical vapor deposition (MOCVD) are reported. The effects of thermal cycle annealing on the crystallinity and the lasing characteristics of GaAs/Si are also reported. By using the AlGaAs/AlGaP ILs, SQWs with a specular surface morphology and a smoother heterointerface can be grown on a Si substrate. Thermal cycle annealing is found to improve the crystallinity of GaAs/Si and to contribute to room-temperature continuous-wave operation of lasers on Si substrates. The combinations of the techniques of AlGaAs/AlGaP ILs and thermal cycle annealing improve the lasing characteristics: an average threshold current density of 1.83 kA/cm2, an average differential quantum efficiency of 52%, an internal quantum efficiency of 83%, an intrinsic mode loss coefficient of 23cm-1, a differential gain coefficient of 1.9cm/A, and a transparency current density of 266 A/cm2, which are superior to those of the two-step-grown laser on a Si substrate. The improvements of the lasing characteristics result from the smooth heterointerfaces of the AlGaAs/AlGaP ILs.

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.

This article briefly looks at the future of telecommunication education in the universities as it evolves from present concerns and trends. Five year bachelor's programs and top-down curricular design will be common. Textbooks supplemented by advance organizers, instruction and testing according to individual learning styles and global integration of education using multi-media services and broadband technology will be some of the other features. Finally, the importance of industry-university partnership in all aspects of engineering education is emphasized.

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.