Abstracts:

In the recent decades technological breakthroughs in science and engineering have led to an explosion in the amount of data available in several fields such as environmental, biological and chemical fields. One of the obstacles preventing this data from empowering new discoveries is the lack of adequate methods that can manage this data and turn it into knowledge. This paper presents a scalable solution to the management of life science databases. Life science web databases are often heterogeneous, geographically distributed and contain semi-structured data. The proposed system (BACIIS: Biological and Chemical Information Integration System) integrates these web databases on-demand. The architecture of BACIIS is decentralized. This design choice was made in order to overcome some of the limitations of remote web-based querying and to create a system that can adapt to an increasing number of users. This paper discusses the architecture of BACIIS and presents an analysis of its performance in response to queries submitted by multiple users. (pp. 3-14)

Keywords: integration, biological databases, distributed architecture

Sequence analysis is widely used to infer function from one protein sequence to another. One of the remaining hurdles in pushing further the limits of efficient usage is the modular and mobile architecture of proteins. Although many resources provide precompiled signatures, they are not yet used in conjunction with pairwise comparisons to investigate the modular architecture and phylogeny. We present a program, doMosaic, which combines existing domain definitions with an adapted sequence alignment algorithm to analyse the possible phylogeny of domain architecture in proteins. It is based on the Smith-Waterman scheme, can be combined with trees derived from the domains and provides a user-friendly graphical interface. The method enables fast and efficient analysis of domain duplication events within one or in-between two sequences. Therefore, it enables refined functional annotation. (pp. 15-20)

Keywords: sequence analysis, domain evolution, data visualisation

Shin-Mu Tseng and Ching-Pin Kao
Institute of Computer Science and Information Engineering, National Cheng Kung University
Tainan, Taiwan, R.O.C.
E-mail: tsengsm@mail.ncku.edu.tw

The microarray technique has been widely used in recent years since it can capture the expressions of thousands of genes in a single experiment. To meet the challenge of high volume and complexity of microarray data, various data mining methods and applications have been proposed for analysing gene expressions. Although numerous clustering methods have been studied, they can not provide automation, high quality and high efficiency simultaneously for the biologists during the analysis process. In this research, we propose an integrated approach that can analyse large volume of gene expression data automatically and efficiently. Our approach integrates efficient clustering algorithms with a novel validation technique such that the quality of the discovered gene expression patterns can be evaluated on the fly. Through practical implementation and applications on real gene expression data, our approach was shown to outperform other methods in terms of efficiency, clustering quality and automation. (pp. 21-28)

Fault diagnosis has become an issue of primary importance in modern process automation as it provides the prerequisites for the task of fault detection. The ability to detect the faults is essential to improve reliability and security of a complex control system. When a physical parameter change due to failure has occurred in a system, the failure effect will hardly be visible in the output performance. Since the failure, effect is reflected as a change in the predictor model. In this paper we describe a completed feasibility study demonstrating the merit of employing hybrid parameter-estimation and fuzzy logic for fault diagnosis. In this scheme, the residual generation is obtained from input-output data process, and identification technique based on ARX model, and the residual evaluation is based on fuzzy logic adaptive threshold method. The proposed fault detection and isolation tool has been tested on a magnetic levitation vehicle system. (pp. 29-38)

Chun-yan Bai, Roberta Houston and Gui-liang Feng
The Center for Advanced Computer Studies, University of Louisiana at Lafayette
Lafayette, LA 70504
E-mail: cxb7146@cacs.louisiana.edu, rah1231@cacs.louisiana.edu, glf@cacs.louisiana.edu

Multicast Re-keying means the establishment of a new session key for the new subgroup in the multicast system. Practical construction methods for multicast re-keying scheme using Reed-Solomon codes and Algebraic-Geometric codes are presented in this paper with examples to show the detailed constructions. The constructions require no computational assumptions. The storage complexity for group members (Group Controller and other users) and the transmission complexity for the schemes have been reduced to O(log(n)) at the same time. (pp. 39-48)

A transformation is proposed which, given a specification of the required external behaviour of a distributed server and a partitioning of the specified service actions among the server components, derives a behaviour of individual components implementing the service. The adopted specification language is close to Basic LOTOS. Unlike in other protocol derivation algorithms based on LOTOS-like languages, distributed conflicts in the given service are allowed, and resolved by self-stabilization of the derived protocol. (pp. 57-74)

Keywords: distributed service implementation, automated protocol derivation, LOTOS

Jen-Chih Lin
Department of Information Management, Jin-Wen Institute of Technology
No. 99, An-Chung Rd., Hsin-Tien City, Taipei, Taiwan, R.O.C.
E-mail: yachih@ms13.hinet.net
AND
Steven K. C. Lo
Department of Computer Science and Information Engineering
Tamkang University, Tamsui, Taiwan, R.O.C.
E-mail: kclo@cs.tku.edu.tw

We develop novel algorithms to facilitate the embedding job when the Flexible Hypercube contains faulty nodes. We present strategies for reconfiguring a complete binary tree into a flexible hypercube with n-expansion. These embedding algorithms show a complete binary tree can be embedded into a faulty flexible hypercube with load 1, congestion 1 and dilation 4 such that O(n^2-m^2) faults can be tolerated, where (n-1) is the dimension of a Flexible Hypercube and (m-1) is the height of a complete binary tree. These methodologies are proven and these algorithms are present to save them. (pp. 75-80)

Keywords: Flexible Hypercube, hypercube, embedding, complete binary tree

Michael J. Pont
Embedded Systems Laboratory, Department of Engineering, University of Leicester
University Road, LEICESTER, LE1 7RH, United Kingdom.
E-mail: M.Pont@le.ac.uk

We are concerned in this paper with the use of “design patterns” to facilitate the development of software for embedded systems. The particular focus is on embedded software with a time-triggered architecture, using co-operative task scheduling. Such “TTCS” software is known to have very predictable behaviour: such a characteristic is highly desirable in many applications, including (but not restricted to) those with safety-related or safety-critical functions. In practice, TTCS archi-tectures are less widely employed than might be expected, not least because care must be taken during the design and implementation of such systems if the theoretically-predicted behaviour is to be obtained. In this paper, we seek to demonstrate that the use of appropriate patterns can greatly simplify the process of creating effective TTCS software. (pp. 81-88)

Keywords: Pattern, Design Pattern, Embedded System, Co-operative, Time-Triggered, Microcontroller

Jean-Claude Royer
Departement d'Informatique de l'Ecole des Mines de Nantes,
4, rue Alfred Kastler. B.P. 20722 F-44307 NANTES Cedex 3
Phone: +33 2 51 85 82 05
Fax: +33 2 51 85 82 49
E-mail: Jean-Claude.Royer@emn.fr

This paper outlines a practical use of algebraic specifications for the development of heterogeneous software systems. This kind of systems mixes several viewpoints, e.g. static, functional and dynamic aspects. Writing, from scratch, an algebraic specification for such systems is quite difficult, so we developed the concept of Graphic Abstract Data Type (GAT). In this paper we present a method to build an algebraic specification of a sequential system \textit{via} a symbolic transition system (STS). The STS models both the dynamic aspects and the static aspects of the system. The STS is also the basis of an algorithm that computes the functional aspects of the system (an algebraic specification). Computing the specification is partly automatic, this improves the compatibility between the aspects. This approach is extended to concurrent and communicating systems by the use of a synchronous product of STSs. We proved that the STS is an abstract interpretation of the generated specification. We demonstrate that the set of axiom may be transformed into a terminating term rewriting system. Then from the generation method of the specification the properties of consistency and completeness are got and this ensures the existence of a partial initial algebra. We showed that the synchronous product of GATs preserves the state predicates, the preconditions and the definedness predicate of the components. We also give sufficient conditions to get the GAT determinism and the GAT compactness of the product of two GATs. (pp. 89-104)

Keywords: Dynamic Behaviour, Algebraic Data Type, Symbolic Transition System, Synchronous Product, Concurrency, Communication, Formal Specification, Mixed System

Dušan M. Kodek and Marjan Krisper
University of Ljubljana, Faculty of Computer and Information Science
Tržaška 25, 1000 Ljubljana, Slovenia
E-mail: duke@fri.uni-lj.si

We consider the problem of optimal allocation of components to a printed circuit board (PCB) assembly line which has several nonidentical placement machines in series. The objective is to achieve the highest production throughput by minimizing the cycle time of the assembly line. This problem can
be formulated as a minimax approximation integer programming model that belongs to the family of scheduling problems. The difficulty lies in the fact that this model is proven to be \emph{NP}-complete. All known algorithms that solve the NP-complete problems are exponential and work only if the number of variables is reasonably small. This particular problem, however, has properties that allow the development of a very efficient type of branch-and-bound based optimal algorithm that works for problems with a practically useful number of variables. (pp. 105-114)

Keywords: combinatorial optimization, integer programming, minimax approximation