ABSTRACTS

Jie Wu and Kejun Yao
Department of Computer Science and Engineering, Florida Atlantic University, Boca Raton, FL 33431

We consider in this paper an optimization model of concurrent software with redundancy. The software fault-tolerance technique used is conversation, a special mechanism for concurrent systems. Two optimization models are combinatorial model and fault-tree model which calculate system reliability based on the same method but under different fault assumptions. These two models are illustrated by examples.(pp. 291-300)

Keywords: concurrent systems, fault tolerance, fault-tree, software reliability

Xiaozhong Li
National Research Center for Intelligent Computing Systems(NCIC), P. O. Box 2704, Beijing 100080. P. R.Chaina
Shuo Bai
National Research Center for Intelligent Computing Systems(NCIC), P. O. Box 2704, Beijing 100080. P. R.Chaina; bai@tango.ncic.ac.cn
Zemin Liu
Beijing University of Posts & Telecommunications

In this paper, we introduce a method of generating control rules for fuzzy logic controllers, called an algorithm of self-learning and self-completing (ASLSC). This method can learn to generaterules from the sample data with the unsupervised competitive learning (UCL) of the neural network. If a rule table generated is not full, the ASLSC method can fill those blanks in the rule table with proper rules. Besides, ASLSC can determine automatically the universes of discourse of fuzzy variables. The simulations on an inverted pendulum system and the experiments on temperature controlling of an industrial electric heating ring show that ASLSC is practicable and effective. (pp. 301-312)

Keywords: self-learning, self-completing, fuzzy control

Jiann-Fu Lin and Sao-Jie Chen
Department of Management, Takming Junior College of Commerce, Taipei, Taiwan, R.O.C.

This paper investigates the problem of scheduling "parallel tasks" with consideration of setup time on a d-dimensional hypercube system, where scheduled tasks are independent such that each task can simultaneously require a subcube for its processing under the constraint that the dimension of a required subcube cannot be greater than the maximum parallelism dimension of a task. Whenever there is a switching of schedule from a task to another one, setup time is necessary. The objective of this problem is to find a schedule with minimum finish time, such a scheduling problem is NP-hard. Therefore, in this paper, we will propose two heuristic algorithms for this kind of problem and derive their performance bounds, respectively.(pp. 313-318)

Keywords: heuristic algorithm, scheduling, parallel task, setup time, performance bounds

Mengchi Liu
Department of Computer Science, University of Regina, Saskatchewan Canada S4S 0A2

This paper proposes a higher-order deductive object-oriented database language called HLO. This language is capable of directly and naturally supporting object-oriented features such as object identity, complex objeCTS, classes, and multiple inheritance. It supports the separation of schema and instance but treats schema as meta data so that schema and inheritance information can be represented and queried in the same way as ordinary data. The language is given a firm logical foundation in the form of the Herbrand-like semantics. The main novelty of the language is the incorporation of the well-definedness and well-typedness constraints in its semantics. A number of interesting semantic properties are investigated. The intended semantics of an HLO program is given by the least and justified model if it can be obtained by a bottom-up least fixpoint computation. The necessary condition to guarantee the existence of the intended semantics and the decidability of checking the necessary condition are also discussed.(pp. 319-332)

Keywords: deductive databases, object-oriented databases, fixpoint semantics, least model semantics

Jiri Slechta
Member of New York Academy of Sciences, 18 Lidgett Hill, Leeds 8, LS8 1PE, U.K.

It is discussed the case of the informational `saser' in the `hot' living systems (T<0), for T < 0_ (the absolute negative zero), and explained why the large quanta of information, like that stored and developed by a genius, or a voluminous new discovery, or a marital love, spread less easily, except for T dserp, 0_, than the smaller quanta, like a random love. A way how to improve this property is to design a sparse set of smaller quanta which span the large quantum (e.g. a kind of teaching the large issue, or policy making about it). It is discussed how to design such a set. In the balance of the equivalence of the various forms of the spanning, a generalized equation of the conservation `energy' figures the relative, and not the absolute, precisions. There is provided a formalized framework of the intercommunication among n persons. On the basis of it there is also given the first exact explanation, in existence, why the `two house' sex is superior to any other form of it, why an offspring of higher animals needs two parents, the theory of measurement of the objective reality, the relation between the theory and experiment, the difference between working of the brain of the people with associative and verbatim memories, and modern and underdeveloped societies, etc. The theory presented here is important for the design of an artificial intelligence (the logic of the real self-organizing systems---(fuzzy) physical logic), and methodology of teaching, the theoretical design of the optimal society, etc.(pp. 333-344)

Keywords: `hot' living systems (T < 0_), informational exchange, informational quanta

Anton P. Zeleznikar
An Active Member of the New York Academy of Sciences, Volariceva ulica 8, SI 61111 Ljubljana, Slovenia

This article deals with problems pertaining to the elements of axiomatics in traditional (mathematical, symbolic, philosophical) logic and to the problems of newly emerging axiomatics in informational logic. Informational axioms can be derived from propositional and predicate axioms and then, particularized and universalized within the informational domain. Traditional axiomatic formulas of the propositional and predicate calculus can become a rebounding cause for the construction of essentially different axioms in general informational theory. It is shown how propositional and predicate axioms and rules can be informationally extended for the needs of the general informational theory.(pp. 345-370)

Keywords: axioms of parallelism, serialism, circularity and spontaneity; catalogue of informational rules; decomposition axioms, decomposition theorem, deduction theorems, implicative axioms, inference rules; informational calculus, predicate calculus, propositional calculus; substitution rule; various axioms and theorems (implicative, conjunctive, disjunctive, equivalent, negatory, etc.)

Ryszard A. Wasniowski
New Mexico Highlands University, Las Vegas, NM, USA; rwasniowski@acm.org

This paper discusses and shows how computation intensive engineering problems can be modeled on parallel-like simulators and computed efficiently on massively parallel computers. Designed with tens of thousands of processing elements, these machines now offer substantially improved computation times, improved cost performance, and allow rapidly reach higher performance levels. Networks of workstations and software packages such as PVM is another attempt to provide a unified framework within which large parallel programs can be developed on a collection of heterogeneous machines, and make easy transition from sequential processing into parallel processing. The ease of developing parallel algorithms for systems identification using gmdh algorithm will be discussed.(pp. 371-378)

Keywords: parallel computing, gmdh algorithm

Zbigniew M. Wojcik and Barbara E. Wojcik
Smart Machines, 13703 Morningbluff Dr., San Antonio, TX 78216, USA.

Elementary features are detected by calculating the number of objects inside partly overlapping windows fixed in an image plane. Each window's contents is processed by a separate processing element (PE) on a SIMD grid or pyramid architecture. Two neighboring feature chunks are merged by adjacent PEs, and the merged feature chunks are joined in each next lth parallel step by every 2lth PE possessing complementary and adjacent feature. The shape recognition terminates if a complete set of features (such as curve bounds, endings, corners, segments, regions forks and junctions) making a whole object meet together on one PE. This complete set is proved to be sufficient to terminate the shape recognition. Hierarchy of elementary features for a merging process is established to recognize object bounds and more primitive features first to have a complete feature set adequate for the consideration by the termination condition at a given level of parallel image recognition (i.e., primal sketch, 2 1/2D sketch and 3D world model). The approach has the property of mapping of image segments directly into phrases and English sentences.(pp. 379-390)

Keywords: parallel shape recognition

Lars Bendix
Institute for Electronic Systems, Aalborg University, Fredrik Bajers Vej 7E, DK- 9220 Aalborg, Denmark; bendix@iesd.auc.dk

After having established the basic terms of the field of software development, we present a conceptual framework to help establish the key tasks to be performed in this field. The field is characterized by two orthogonal concepts: Programming-in-the-Large (PitL) and Programming-in-the-Many (PitM). These concepts can be further subdivided into the tasks of Configuration Management and Version Control (for PitL), and Personnel Management and Resource Management (for PitM). The main body of this paper is dedicated to a thorough analysis of these tasks. The conceptual framework is then applied to these systems to show how to use it to evaluate and compare systems. Finally is given a discussion of the presented framework, both in its own rightand with relation to other work.(pp. 391 - 406)

Keywords: software development environments, taxonomy, configuration management, version control, personnel management, resource control

Mirko Malekovic
University of Zagreb, Faculty of Organization and Informatics, Pavlinska 2, 42000 Varazdin, Croatia

A formal system for functional dependencies is introduced. The formal system contains only two inference rules: (I) Reflexivity and (II) Generalized transitivity.(pp. 407 - 418)

Keywords: completeness, formal systems, functional dependencies, inference rules, soundness