Advanced Search

1998 Vol. 24, No. 4

  Previous Issue | Next Issue 
Display Method:
Inductive Inference and Inductive Process
Li Wei
1998, 24(4): 373-381.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 903
Abstract:
This paper analyzes the difference between inductive inference and deductive inference, presents a formal logic framework for inductive inference, and defines the concepts of reconstruction and inductive sequence. In addition, it gives an inductive procedure to generate inductive sequences. It proves that for a model M, if the set εM of all its instances is known, then starts with any theory, all inductive sequences generated by the procedure will converge to the same limit, i.e., the set of all true sentences of model M. Therefore, it justifies the inductive rules.
Knowledge-Base Refinement Mechanism towards Machine Translation
Guo Honglei, Li Wei
1998, 24(4): 382-385.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 971
Abstract:
Some principles to refine knowledge base are proposed. A knowledge-base (KB) refinement mechanism towards machine translation is provided,in which, the importance of errors serve as the refining indicators and the most serious errors are solved first, and the effectiveness and acceptance of KB refinement are determined based on system performance and the KB convergent gain alternation. The error identification strategy via case-based reasoning is presented. Generalizing and specializing operations based on focus revision are presented as well.
Fractal Image Compression on Wavelet Domain
Li Bo, Lei Zhenyu
1998, 24(4): 386-390.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 965
Abstract:
After analyzing the cause for generating tiling effect, an objective measure for calculating tiling effect is defined. In order to eliminate tiling effect, an original image is first transformed into wavelet domain, on which fractal compression and arithmetic coding are performed. This method inherits the principles of fractal coding and exploits the good characteristics of wavelet transform. Furthermore, similarity matching between blocks is sped up by utilizing the energy distribution among wavelet coefficients. As a result, image compression performances are improved in both PSNR and coding speed.
Confidential Theory of Inheritance in Nonmonotonic Semantic Network
Ma Dianfu
1998, 24(4): 391-394.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 709
Abstract:
The confidential theory describes a new approach to inheritance reasoning in nonmonotonic semantic network allowing for multiple inheritance with exceptions. The approach leads to an analysis of defeasible inheritance that is well-defined and intuitively attractive. It yields unambiguous results applied to any acyclic semantic network.Etherington's default logic of inheritance is discussed. It provides a new confidential default logic of inheritance,describes properties of confidential default logic which has unique extension. The process of strutting unique extension is recurrence. The extension of confidential theory can compute quickly. These fast inference algorithms are applicable to expert system.
Implementation of an Analogical Reasoning Coprocessor
Shen Xukun, Zhao Qinping
1998, 24(4): 395-398.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 856
Abstract:
Analogical reasoning technique is significant to the development of AI(Artificial Intelligence) systems . Computing systems based on conventional architecture are very inefficient in doing analogical reasoning,which obstructs the research and application of analogical reasoning. To solve this problem, a high-performance embedded analogical reasoning system called ARC(Analogical Reasoning Coprocessor) is presented. Since ARC has a specialized architecture with special speedup strategy to support analogical reasoning mechanism, and is implemented with advanced integrated devices.It is much faster than conventional computing systems in doing analogical reasoning.
Research and Implementation of Data Warehouse and Data Mining in Decision Support Systems
Wang Bingqing, Huai Jinpeng
1998, 24(4): 399-402.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1137
Abstract:
DSS (Decision Support System) is to support decision makers and improve the decision-making process,which is much promising in CIMS,virtual enterprise,medical diagnosis and so on.Some techniques related to DSS are introduced,such as data warehouse,knowledge discovery in database and distributed object.Then a kind of integrated IDSS development environment is put forward, which combines the traditional IDSS with the data warehouse, applies a special model named data mining model to knowledge discovery in database and adopts distributed object technique in implementation. Finally,the design and the implementation of the data warehouse and knowledge discovery in database are introduced.
Software Architecture Model of Enterprise Information System Domain
Tao Wei, Chen Chong, Mai Zhongfan
1998, 24(4): 403-406.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 812
Abstract:
Software architecture is a new method in Software Engineering.It provides a higher framework for software reuse. This paper introduces the basic concept and research achievement of software architecture into the construction of enterprise information system and puts forward the software architecture model.It also gives a survey on the software engineering methodology and elucidates the significance of software architecture in context of software process. Then, a software architecture based model of enterprise information system domain is introduced,including software architecture model, reference requirement, reference architecture and so on. Thirdly, the software development process based on the model is described. Finally, some relative researches are discussed.
Research on UML Software Development Process and Support Environment
Zhang Li, Ge Ke, Wang Yun, Zhou Bosheng
1998, 24(4): 407-410.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1433
Abstract:
UML (Unified Modeling Language) is one of the latest fruits in software engineering field. It terminated the fighting in the field of modeling language of object oriented technology. But how to use UML to solve the real world problem is still the hot topic on the field. From the view of the requirements for software development, the flexible software development process is presented to fit the changing world, and the framework of integrated UML software development environment is designed to support modeling, simulation and round-trip development. The environment should be composed of UML visual modeling system,UML simulation system,UML code generation system and UML reverse translation system.
Reversed Automatic Generation of Visualized Class Diagram of Object-Oriented Program
Liu Chao, Li Jian, Shen Haihua
1998, 24(4): 411-414.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1013
Abstract:
Reversed modeling is the critical technique in analyzing and understanding object-oriented software, which is also the basis of its testing and maintenance.A set of basic rules for reversedly constructing program's class models are studied, and an inheritance-layered class diagram is proposed, which can be drawn automatically. Kinds of relationship, such as inheritance, composition and association, can be shown in the layered class diagram. By the diagram, relativity and dependency among classes can be simply highlighted. In the end, the revised modeling tool, SafePro/Re-OM, based on the set of rules, is presented.
Study on Hierarchical Model for Ada Parallel Softwares
Yao Shuzhen, Tang Fagen, Jin Maozhong
1998, 24(4): 415-417.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 928
Abstract:
Having visualization and formalization, Petri net has been used to model and analyze parallel softwares. Comparatively,OPN(Object-oriented Petri Nets) is more capable of modeling.A hierarchical parallel software model based on OPN is presented. It refers to UML(Unified Modeling Language) description in its structure, describes the internal state of an object and mutual interaction among objects by means of ECPN, making simulation of parallel softwares and translation to Ada programs easier. The model will be useful for quality assurance of parallel softwares, specially Ada95 parallel softwares, and improvement of development efficiency.
Design and Implementation of Modelbase and Algorithmbase Based on the Component Object Model
Wang Baojiang, Huai Jinpeng, Xia Wanqiang
1998, 24(4): 418-421.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1383
Abstract:
This paper provides a kind of design theory and system structre of the intelligent decision support system development enviroment in the analysis and design process of this development enviroment implementation. It also discusses the design and implementation of modelbase and algorithmbase in decision support system, gives the analysis, design and implementation of modelbase and algorithmbase based on the object modeling technology and component object model, on the basis of the analysis of object model technology and distributed object technology.
Integrated Process Engineering Environment and its Architecture
Wang Lei, Zhou Yuchen, Gao Zhongyi, Lu Boying
1998, 24(4): 422-425.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 791
Abstract:
Modern enterprise is a very complicated integrated human-machine system, of which internal enactment needs the supports by computer-based application system. So a process-centered and computer-aided integrated process engineering environment combining process engineering and software engineering together is discussed, which supports the whole life cycle of an enterprise process, the development of computer-base application system,the production and management of real product within enterprise. A whole process model based framework is presented, and a distributed architecture of process enactment environment is designed.
Empirical Study of Personal Software Process
Che Xiangdong, Xu Hong, Ma Yunjing, Liu Youcheng
1998, 24(4): 426-429.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1002
Abstract:
To verify the effect of the PSP(Personal Software Process) on software development,Software Engineering Institute of Beijing University of Aeronautics and Astronautics made an experiment on the application of the PSP in 1997.After a brief introduction to the principles and methods of the PSP,the experiment is presented, its data are given out, and these data are then analyzed in terms of development productivity and defect density.Result of the analysis shows that the application of PSP can significantly improve the productivity of software development and the quality of the product.In addition, problems met in the experiment and their causes are also discussed, and further research and work that should be done are pointed out.
MIS with Visual System Modeling Language
Tan Wenan Zhang Xiuguo, Xu Hong, Lu Boying, Zhou Bosheng
1998, 24(4): 430-433.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 771
Abstract:
VSML(Visual System Modeling Language) is successfully used for describing computer-based application systems from functions, behaviors, data and architecture, and supporting object-oriented rapid prototyping and software reuse methods.VSML's design considerations, the definition of primitives, the behavior description rules and the improvement of database access functions are firstly introduced, then its architecture and application in the development about MIS are described in detail. Finally, a conclusion and thought for further study about VSML are given out.
Genetic Algorithms and Its Application in Software Test Data Generation
Jie Wei, Xie Junkai, Xi Hongyu, Gao Zhongyi
1998, 24(4): 434-437.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1972
Abstract:
Test data generation is a hard problem in software testing area and its implementation can greatly enhance the automation of software testing. The main contribution of this paper lies in that a pratical model, which utilizes genetic algorithms as searching policy to generate software structural test data , is proposed. To achieve higher performance, such issues as coding strategy, evaluation function construction and instrumentation are addressed in detail. In particular, much emphasis is put on algorithms operator evolution because the employ of operator is a key factor which can highly affect algorithms efficiency. Finally, some convincing experiment results are presented which show that our genetic algorithms-based method is more efficient comparing with other methods
Encoding Methods of Parameters of Genetic AlgorithmsApplied in Ada Dynamic Testing
Shi Xiaohua, Gao Zhongyi
1998, 24(4): 438-440.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1035
Abstract:
The principles and methods are introduced, which are used to encode parameters when applying the genetic algorithms in Ada structural testing. First , two main principles of encoding are introduced. Because Ada language could use dynamic data types in its parameter tables, the method used to handle them would be the most important problem when we want an automatic dynamic testing tool . Secondly , encoding methods used in both enumerative data types and dynamic data types are introduced. Parameters of enumerative data types have the same property. They can be encoded in the same way of integers. Encoding methods of the parameters of dynamic data types are emphasized.
Practical Metric for Software Testing
Deng Yi, He Zhitao
1998, 24(4): 441-444.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1186
Abstract:
The measurement to software testing process is addressed.Software testing is to find software internal problems. It is crucial part of quality assurance. But it is pity that we cannot achieve complete testing due to practical limitations and theoretical impossibility. And just as the old saying "fail to plan, plan to fail",good testing requires thinking out a test plan, defining the responsibilities and tasks during the development of product. They are the essential prerequisites to prepare for and manage the testing in the realworld. Software Metric for software testing is the measurement of quantity that has impact on software testing. This paper makes a brief introduction to software testing measurement and then introduces a practical metric for the schedule and software problems counting in software testing.
Debugger Design of Fault Tolerant Computer
Jin Huihua, Liu Yanpeng, Chen Yongping
1998, 24(4): 445-448.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 838
Abstract:
The debugging method and tools of fault tolerant computing systems are explored. The presented approach is based on use of RT scope multitask debugging tools at the operating system level on a prototype of the system considered. The paper first discusses requirement on environment during developing fault tolerant computer. The system architecture of debugger for fault tolerant computer is presented. The driver is designed to perform real time communicationn function.The cross debugging environment of real-time multitasks is set up under Windows. The results show that it greatly decreases the complexity of computer system and brings much usefulness into developing and testing.
Improvement and Implementation of Marching Cubes Algorithm Based on Surface-Based Rending
Shen Xukun, Liu Jingwei, Zhao Qinping
1998, 24(4): 449-453.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 952
Abstract:
Three-dimensional imaging of medical objects from volume data is considered increasingly useful in various fields, such as medical diagnosis, forensic medicine, palaeoanthropology etc.Marching Cubes algorithm, as one of the advanced surface-based rendering algorithms,is analyzed.The algorithm is improved in several aspects, including solution in several ambiguous cases and shortage of memory, reduction of the computing cost.An advanced 3D medical volume data rendering prototype system is provided, which is based on SGI Indigo2 Open GL platform and consists of a powerful display engine and has satisfying solution.
Design and Implementation of High Performance BLAS for Pentium Pro
Li Zhongze, Chen Jin, Long Xiang, Li Wei
1998, 24(4): 454-457.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 850
Abstract:
BLAS is basic linear algebra libraries for science and engineering applications.It plays an important role in high-performance computing. By identifying and optimizing frequently used, numerically intensive operations, BLAS can help in reducing the cost of computation, enhancing portability, and improving productivity.By taking advantage of the architectural features of Pentium Pro, a series of optimized methods are proposed to implement BLAS on Pentium Pro so that BLAS can perform optimally on Pentium Pro systems. The testing results demonstrate that the speed of BLAS3 can reach 112Mflops on 200MHz Pentium Pro.
Design of the Link Layer Protocol of the High Performance Cluster Network
Wu Wenjun, Xiang Xiaohua, Long Xiang
1998, 24(4): 458-461.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 967
Abstract:
A high performance computer cluster system requires an interconnect network with high bandwidth , low latency and high reliability. Only when a simple and effient network protocol is developed , a high performance interconnect network will come into being. A data link layer protocol framework suitable for computer cluster network is provided.The stop-and-wait protocol and the go back N protocol are discussed . FPGA(Field Programmable Gate Array) chips are used to implement their logic separately and their performances are analyzed.The result of the analysis shows that go back N whose implementation requires almost the same FPGA resources as that of stop-and-wait can achieve better performance than it.
Unified Parallel Debugged Object Model for PVM
Huang Ning, Jin Maozhong
1998, 24(4): 462-465.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 861
Abstract:
It is important to well understand the debugged parallel program in researching and developing its parallel debuggers.Although many parallel debuggers define events to be occurrence instances of program actions,little of them give the complete definition of action. PVM(Parallel Virtual Machine) system is analyzed, and concrete actions are defined,which are used to describe a unified PVM debugged object model for PVM.A same debugged object can be provided in all debugging periods when using this model,this makes parallel debugging more easy.
Hierarchical Network Management System
Sun Yaqiu, Zhang Weidang, Li Weiqin
1998, 24(4): 466-469.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 762
Abstract:
A hierarchical network management system is presented,which consists of a main manager and many submanagers. The main manager delegates management tasks to submanagers by using NMP(Network Management Procedure). Idealy, the main manager runs at the management center and the submanager runs at the managed network site, hence it is a hierachical architecture.The NMPs excute in the submanager, communicate with agents staying in the managed device, gather data and process these data, then return the processed data back to the main manager,thus reduces the management traffic.
Firewall System with Strong Authentication
Zhu Chun, Yan Jihe, Li Weiqin
1998, 24(4): 470-474.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 923
Abstract:
The fundamental theory,key technology and prevailing solution in the security fields were analyzed.A new firewall architecture was discussed in detail.Compared with traditional firewall,it has markedly advantage in generality,transparency and authentication.Firewall,as a comprehensive security protection system,does not mean a single technology or product.Actually,its functionality is the result of cooperation of all basic security components.Particularly,authentication,as the basis of other security function,has specfic importance.We designed and developed a practical security protection system suitable to deployment on the edge of local campus network,which combined the advanced security characteristic of strong authentication mechanism and general proxy firewall architecture.
Study on CAD/CAE System Architecture Supporting MEMS
Kang Jianchu, Yin Baolin, Gao Peng
1998, 24(4): 475-478.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1016
Abstract:
CAD/CAE technique plays an important role in MEMS(Micro Electro Mechanical Systems) research, while MEMS CAD/CAE shares some common techniques with conventional CAD/CAE systems, it is also different in many ways.The characteristics of CAD/CAE techniques supporting MEMS were introduced firstly, then the system architecture of MEMS CAD/CAE was studied, the structure frame of software supporting tool was presented and the ways for solving the key techniques were expounded. The application of CAD/CAE technique will improve the design quality and shorten the research time of micro electro-mechanical systems. As a result, it will lead to industrialization of MEMS sooner.
Content-Based Visual Multimedia Query Language
Gong Shengrong, Xiong Zhang, Liu Xudong
1998, 24(4): 479-482.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 897
Abstract:
At present, there are two ways to get multimedia information from multimedia database. One is applying query language to manipulate information directly,it can exactly tell computer about relative query information,but its user interface is not friendly.The other one is developing query interface,but it is lowly precise.With the development of visual programm language enviroment and the technique of content-based query,it is possible to combine both of them. Therefore,query by visual example and query by pictorial example are connected,and a content-based visual multimedia query language—CBVMQL is designed,in which semantics expression, concept definition, object-oriented expression are combined. CBVMQL not only provides a visual query interface,but supports manipulation exactly of information by automatic query.Otherwise,it implements content-based multimedia query,too.
Description and Presentation of Multimedia courseware
Liu Xiaodong, Jin Maozhong
1998, 24(4): 483-486.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 811
Abstract:
The Key problem of developing multimedia courseware authoring system is how to present the instruction flow and presentation.GNT(Generalized Transition Network) is used to describe instruction flow and instruction content is expressed by PC(Presentation Context)in multimedia courseware.In the meantime,how to use them in BUAA-MCAS(BUAA Multimedia Courseware Authoring System) is discussed. It includes how to describe instruction flow with GTN and how to present instruction contents with presentation context.The internal data structure is also given.These methods in BUAA-MCAS have two characteristics.First,the description of instruction flow and presentation of instruction contents are separated,which makes the courseware authoring system flexible.Secondly,instruction contents can be concurrently presented by expressing the instruction presentation with presentation contexts.
Research of Simulation Technique for Concurrent Software on OPN
Tang Fagen, Yao Shuzhen, Deng Yanqun
1998, 24(4): 487-490.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1080
Abstract:
The performance evaluation for concurrent software is one of the important means to improve the quality of development. This paper mainly discusses applying the combination of object-orientation and Petri nets to the system modeling of concurrent software development, and the simulation technique for concurrent software based on OPN(Object Petri Nets). It also discusses the key techniques and solutions when implementing a dynamic simulation tool based on OPN.The initial experiment results show that this implementation mechanism and processing methods discussed here are effective and feasible.
Object-Oriented Petri Net Based Method for the Concurrent Program Modeling
Ren Aihua, Niu Jinzhong, Zhang Yongming
1998, 24(4): 491-494.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 1497
Abstract:
An approach for concurrent system modeling based on the object-oriented high-level Petri net is described. The O-O Petri net is constructed by combining object-oriented techniques with Petri net theory. The modeling of the O-O Petri net not only can reduce the complexity of the model, but also make the modeled system with reusability and easy maintainability. This approach makes O-O system modeling with the description for both formalization and graphics.As an illustration,the producer and consumer concurrent pattern is modeled using object-oriented Petri nets.The modeling experience with these nets demonstrates that the decoupling and separation of knowledge and constraints clearly enhance maintenance and reusability in concurrent system modeling.
Algorithm for Constructing the Simplified DFA of Regular Expressions
Tan Fengqin
1998, 24(4): 495-498.
Abstract References Relative Articles Resource File Access Statistics iconViews $!{v.htmlCount} iconDownload 3029
Abstract:
An algorithm for constructing the simplified DFA(Deterministic Finite Automaton) is introduced,which is equivalent to a given regular expression. The first step constructs an NFA(Nondeterministic Finite Automaton) equivalent to the regular expression, where the operation for constructing finite automata with ε-moves is omitted. Then the simplified DFA equivalent to the NFA is constructed by using state transition trees. The algorithm bas been realized by computer programming, and for any input regular expression, a simplified DFA equivalent to the regular expression is produced. The algorithm can be applied in the design and analysis of the system with discrete inputs and outputs.

Copyright © Journal of Beijing University of Aeronautics and Astronautics

Address: Editorial Department of Journal of Beijing University of Aeronautics and Astronautics, 37 Xueyuan Road, Haidian District, Beijing Post Code: 100191 Email: jbuaa@buaa.edu.cn

Supported by: Beijing Renhe Information Technology Co., Ltd.