Showing papers on "Design paradigm published in 1992"

Assuring Design Diversity in N-Version Software: A Design Paradigm for N-Version Programming

Abstract: The N-Version Programming (NVP) approach achieves fault-tolerant software units, called N-version Software (NVS) units, through the development and use of software diversity. To maximize the effectiveness of the NVP approach, the probability of similar errors that coincide at the NVS decision points should be reduced to the lowest possible value. Design diversity is potentially an effective method to get this result. It has been the major concern of this paper to formulate a set of rigorous guidelines, or a design paradigm for the investigation and implementation of design diversity in building NVS units for practical applications. This effort includes the description of a most recent formulation of the NVP design paradigm, which integrates the knowledge and experience obtained from fault-tolerant system design with software engineering techniques, and the application of this design paradigm to a real-world project for an extensive evaluation. Some limitations of the approach are also presented.

Unifying multi-paradigms in software system design

Abstract: Large software systems are often aimed to deal with problems of multi-application domains or disciplines, where each class of problems has its own defining features, difficulties and emphases. No existing computation model or design paradigm is appropriate for all varieties of problems faced in software design. A unification of existing computation paradigms is thus necessary to accommodate the diverse needs in system design. In this paper, the authors propose a framework based on the notions of logical entity abstraction and instantiation (LEAI), which enables a user to select appropriate computation paradigms to design various parts (or subsystems) of a software system, and to compose the parts into a integrated system. An example is given to illustrate the application of the framework. >

Complex behavior specification for autonomous systems

Abstract: The author focuses on one of the problems of design of an autonomous reactive system, namely, specification of its behavior with respect to often unpredictable changes occurring in the real world. Several behavior-oriented languages are briefly presented, and some formal relationships between two of them, process transition networks and statecharts, are given. The goal is to address the need for augmenting the behavior-based design paradigm with theoretical tools adequate for expressing complex behaviors of autonomous reactive systems pursuing high-level goals and applicable to both analysis and synthesis tasks for such systems. They also provide some basis and suggestions for further research on unifying classical and artificial-intelligence-based approach to this problem. >

Herbert Simon's 'Science of Design': two decades later

Abstract: The Karl Taylor Compton lectures delivered by H. Simon at the Massachusetts Institute of Technology in 1968 and published the year later as The Sciences of the Artificial (MIT Press, 1969) forms the starting point for the discussion. One of the chapters in Simon's book is titled 'The Science of Design: Creating the Artificial'. It is with the nature of this 'science' that the author is concerned. The author discusses the AI design paradigm, designs as satisficing solutions, bounded rationality, and the operational model of the design process.

The Present and Future of Intelligent Tutoring Systems

Abstract: The goals of this chapter are twofold. First, we will show how the use of machine learning techniques can greatly improve the dynamic construction and updating of student models. Then, we will discuss the underlying principles of the traditional design paradigm for Intelligent Tutoring Systems (ITSs) and will argue that they are, partially, responsible for the difficulties felt by present day ITSs. We maintain the thesis that the teaching (and learning) situation is a particular instance of the general case of interaction among intelligent agents. As a consequence of this point of view, we sustain that new principles for building ITSs are needed, based on the idea of belief systems.

Metaparadigm's adventures: Incorporating design situations in software process design

Abstract: Our practical experiences and theoretical research in the field of software design and its management have resulted in the conclusion that to enable successful software system design in any design situation we must first design an appropriate design paradigm This activity was called the metadesign and a metadesign paradigm a metaparadigm Following recent scientific findings we decided that a metaparadigm should be pluralistic, useful, systemic, easy to use and recursivelly adaptable (recursive adaptability states that a metaparadigm must be an iterative learning system accumulating new knowledge about metadesign, software system design and related) It should be composed out of a suitable framework, theory and a metadesign methodology and philosophy Currently we have designated the idea of a process as the framework, the Checklands Soft System Methodology as the methodology and philosophy, and the process formalization (metamodeling) as the theory

High level design: a design vision for the 90's

Abstract: Summary form only given. The amount of electrical engineering knowledge required in semiconductor design is decreasing and the process is now mimicking software design. Design re-use is moving upstream with sophisticated subsystems and flexible, technology-independent building blocks becoming available from ASIC vendors and other sources. The time spent debugging the applications software is vastly exceeding the development time invested in the hardware. Driven by the opportunity of million-plus gates ASICs on the one hand, sophisticated synthesis and simulation tools on the other, the design paradigm for most designers is shifting from the gate level to the functional level. With this shift comes the move to hardware description languages, increased design re-use, and active trade-off between software and hardware implementations. The author highlights the status and progress of this evolution and discusses its impact on semiconductor and system design in the coming years. >

Modeling the DARPA Diesel Engine in ProEngineer

Abstract: ThE'se Ilotes summarize my experience modeling all moving parts of the DARPA dipsel f"ngine wit.h Pro/Ellgineer 8.0. The purpose of the execCi.9E' has bpf'll to investigatf' t.he suitabilit.y and effectiveness of the design paradigm of an industrial modeling 5yst,em.

A design method for low complexity control laws

Abstract: A design method is presented which leads to low-order control law designs in a natural manner without recourse to explicit model order reduction methods. The method employs parameter optimization to determine compensator coefficients in a framework based on classical SISO design intuition, yet supports the MIMO loop shaping design paradigm, thus providing a measure of robustness to unstructured uncertainties. Two nontrivial examples are included to illustrate the salient features of the method.

A New Architecture Design Paradigm for Parallel Computing in Scheme

Abstract: This paper describes a new architecture design paradigm that radically reassigns various system responsibilities among the compiler, operating system, and architecture in order to simplify the design and increase the performance of parallel computing systems. Implementation techniques for latently typed languages like Scheme are enhanced and used to support compiler-enforced memory protection and compiler-controlled exception handling. Hardware design complexity is greatly reduced and hardware modularity is increased by not only eliminating the need to implement exception handling in the processor state machine, but also by eliminating global control altogether. In the absence of global control, techniques such as pipelining and multiple contexts that exploit instruction-level and thread-level parallelism can be used together without the usual processor complexity problems, to increase the efficiency of parallel systems. Complexity is reduced and efficiency is increased at the software level as well. The use of compiler-enforced memory protection and a single shared system-wide virtual address space increases inter-thread communication efficiency as well as inter-thread protection resulting in threads that not only are light-weight but also enjoy the protection guarantees of heavy-weight threads.