Abstract:This paper focuses on the problem of data representation and explanation of three-dimensional volume data field based on the physics model of the data．First the data model is set up by three-dimensional geometry model．Then the surfaces are extracted from the data field and match with the model to reconstructure the data field based on the Dhysics model．This paper also presents a subdivision algorithm to fast trace B-spline curve and the anti-approach iS adopted to evaluate the control polygon ot the curve in order to extract the geometry feature．

Abstract:Loops is a kind of structure in which the parallelism is the most abundant, so it is the most mportant source of parallelism. However procedure calls in loops obstruct greatly the data dependence analysis in loops and make the potential plentiful parallelism in loops can not be developed. In this paper, a method called loop embedding is presented.Loop embedding can make parallelizing the partial of DO loops with procedure calls become possible. For the partial of DO loops with procedure calls, in other interprocedural data dependence analysis technologies, it's parallelism can be developed, but in loop embedding, the cost is less. In loop embedding, the overhead, which calling the same procedure many times leads to, can be decreased. The scheme introduced in this paper was implemented in AFT (automatic Fortran transformer), a parallelizing compiler the authors developed. The test result on Spec92 illustrates the effectiveness of their method.

Abstract:LOTOS(language of temporal ordering specification) is a protocol specification language based on process algebra CCS. It is geared to protocol verification, but it is not powerful enough for describing some properties of practical protocols. This paper introduces a language ELOTOS(extended LOTOS), with the power of describing asynchronous communication, time, and stochastic event occurring.

Abstract:In this paper，an AOP(agent oriented programming)framework is described from the view of the SO-called cooperative agent application problem．In the framework，a new agent language(AOPL)is presented，the design and implementation of its program-ming system called AOPS are discussed．AOPS can translate AOPL to C．Moreover，a new agent relation model is presented．Detailed discussion is the composition of the model and the importance on embodying the architecture of the cooperative agent application sys-tem，guiding the cooperative behavior among agents，and supporting the in-depth appre-hension about the characteristics of the system architecture．Finally，the application of AOP in the multiple functional perception system which is nOW being implemented is dis-cussed.

Abstract:In this paper，an AOP(agent oriented programming)framework is described from the view of the SO-called cooperative agent application problem．In the framework，a new agent language(AOPL)is presented，the design and implementation of its program-ming system called AOPS are discussed．AOPS can translate AOPL to C．Moreover，a new agent relation model is presented．Detailed discussion is the composition of the model and the importance on embodying the architecture of the cooperative agent application sys-tem，guiding the cooperative behavior among agents，and supporting the in-depth appre-hension about the characteristics of the system architecture．Finally，the application of AOP in the multiple functional perception system which is nOW being implemented is dis-cussed.

Abstract:In this paper，an AOP(agent oriented programming)framework is described from the view of the SO-called cooperative agent application problem．In the framework，a new agent language(AOPL)is presented，the design and implementation of its program-ming system called AOPS are discussed．AOPS can translate AOPL to C．Moreover，a new agent relation model is presented．Detailed discussion is the composition of the model and the importance on embodying the architecture of the cooperative agent application sys-tem，guiding the cooperative behavior among agents，and supporting the in-depth appre-hension about the characteristics of the system architecture．Finally，the application of AOP in the multiple functional perception system which is nOW being implemented is dis-cussed.

Abstract:This paper briefly introduces the technology of agent-based software engineering, and then designs a SAIS(software agent interopration system) for maintaining knowledge bases by using this technology. The SAIS system consists of multiple problem solvers, and each problem solver is a software agent. This paper also shows the structure of the SAIS system, and describes the technical aspects about the implementation of the SAIS system.

Abstract:The authors discuss concisely, the process of high level synthesis and the extraction and transformation of control information generated in the said process. The authors also discuss the methodology of linking the result of data flow synthesis and that of control flow synthesis as well as the mapping to technology-dependent ASIC based on Xilinx FPGAs.

Abstract:In this paper, the scope of discovery of sequential pattern has been extended from single level to multiple level, that is, discover sequential pattern among the objects in the same concept level or in different level. And a top-down and progressively deepening method for discovering sequential pattern in each different level has been put forward.

Abstract:This paper aims at parallelizing flowfield computation applications automatically. Flowfield data and serial flowfield computation programs are partitioned evenly in SPMD model. Arrays shared variables are introduced and the recursive function that automatically searches for arrays shared variables is discussed in detail. An automatic parallelizer has been implemented preliminarily to support the approach in this paper.

Abstract:The multi-concept acquisition algorithm HMCAP and the neural network algorithm FTART are integrated in hybrid learning model HLM，which can deal with multiple concepts and continuous attributes．In this paper，the incremental learning algorithin of HLM which based on the structure of hybrid binary decision tree and FTART network is proposed．It has the ability of adjusting old structure to improve learning accuracy by once learning instead of rebuilding the decision tree and the neural networks when the new examples were provided．