Sunday, August 20, 2017
   
Text Size
Login

Journal of Information and Communication Technology (JICT) Vol. 1, No. 2 Dec 2002

DECISION MAKING USING MODIFIED S-CURVE MEMBERSHIP FUNCTION IN FUZZY LINEAR PROGRAMMING PROBLEM
1Pandian, M. Vasant, 2Nagarajan, R and 3Sazali Yaacob
1Mara University of Technology, 88997 Kota Kinabalu, Sabah
2&3Universiti Malaysia Sabah, 88999 Kota Kinabalu, Sabah
 
 
ABSTRACT FULL TEXT
In order to develop approaches to solve a fuzzy linear programming problem, it is necessary to study first the formulation of membership functions and then the methodology for applying the solution to real life problems. A S-curve membership function is proposed in this paper. It is important to note that S-curve membership function has to be flexible to describe the fuzziness in the problem. Fuzziness may occur in several levels of an industrial production management such as manpower requirements, resource availability such as software and the demand to be met. In order to show that the S-curve membership function works well for fuzzy problems, a numerical example is demonstrated. A thorough study on how the non linear membership function used in dealing with fuzzy parameters and fuzzy constraints is also presented. Only one case where all three coefficients (such as objective coefficients, technical coefficients and resource variables) that normally occur in production planning problem, are considered and fuzzified. However, there are several other cases. The result obtained from this paper is to provide confidence in using the proposed S-curve membership function in a real life production planning industrial problem.
 
Keywords: S-curve function, vagueness, fuzzy parameters and degree of satisfaction.
 

 
EXTENDING THE DECOMPOSITION ALGORITHM FOR SUPPORT VECTOR MACHINES TRAINING
1N. M. Zaki, 2S. Deris and 3K. K. Chin
1&2Faculty of Computer Science & Information System, University Technology Malaysia, Johor, Malaysia
3Engineering Department, Cambridge University, Trumpington Street, Cambridge, UK
 
ABSTRACT FULL TEXT
The Support Vector Machine (SVM) is found to be a capable learning machine. It has the ability to handle difficult pattern recognition tasks such as speech recognition, and has demonstrated reasonable performance. The formulation in a SVM is elegant in that it is simplified to a convex Quadratic Programming (QP) problem. Theoretically the training is guaranteed to converge to a global optimal. The training of SVM is not as straightforward as it seems. Numerical problems will cause the training to give non-optimal decision boundaries. Using a conventional optimizer to train SVM is not the ideal solution. One can design a dedicated optimizer that will take full advantage of the specific nature of the QP problem in SVM training. The decomposition algorithm developed by Osuna et al. (1997a) reduces the training cost to an acceptable level. In this paper we have analyzed and developed an extension to Osuna’s method in order to achieve better performance. The modified method can be used to solve the training of practical SVMs, in which the training might not otherwise converge.
 
Keywords: Support vector machines, decomposition, pattern recognition, and learning.
 

 
OBJECT ORIENTED ANALYSIS AND THE DESIGN OF LARGE CLIENT SERVER APPLICATION IN A WINDOWS ENVIRONMENT: AN EXPERIENCE
1K.Z Zamli, 2W.A Wan-Hassan and 3N.M Mohd-Zainuddin
1Pusat Pengajian Elektrik Elektronik, Universiti Sains Malaysia, Nibong Tebal, 14300 Pulau Pinang
2Consolidated Cable (M) Sdn Bhd, Lot 32, Lebuh Sultan Mohamed 1, Bandar Sultan Suleiman, 42000 Port Klang, Selangor
3Program Pengajian Diploma, Universiti Teknologi Malaysia, Jalan Semarak, 54100 Kuala Lumpur
 
ABSTRACT FULL TEXT
A Unified Modeling Language (UML) is probably the most popular language and notations for Object Oriented Analysis and Design (OOAD) in the industry. In fact, the UML, a unification of James Rumbaugh’s Object Modeling Techniques (OMT), Grady Booch’s Booch Techniques, and Ivar Jacobson’s Object Oriented Software Engineering (OOSE), is fast becoming a lingua franca for software engineers, developers and designers alike. Being a lingua franca, the UML helps software engineers “speak” in the same language. In effect, the UML facilitates reuse of not only codes, but also software architectural designs. In some cases, these architectural designs are also documented as reusable designs or patterns.
 
This paper, derived from our previous work (Idris et al., 2000; Zamli et al., 1999a; Zamli et al., 1999b; Zamli et al., 1999c; Zamli et al., 1999d; Zamli et al., 1999e), describes our experience using a UML to design large scale object oriented client server database applications in a Windows environment. In doing so, we have developed some reusable designs and conventions in terms of UML class diagrams along with class relationships, cardinality and stereotypes, as well as in terms of component diagrams and their dependencies. Using such designs and conventions, we have incrementally developed a Financial Analysis Module as part of a larger Enterprise Resource Planning Systems using the Borland C++ Builder 4.0, Microsoft SQL Server 7.0 and Rational Rose 98i, in a Windows NT platform with an average of 16,600 lines of codes (LOC) and 98 objects.
 
While some aspects of the designs and conventions used in the Financial Analysis Module are project specific (e.g. using case diagrams, collaboration diagrams, and sequence diagrams), our experiences indicate that some aspects of the designs can be applicable in other development projects in a similar context (i.e. involving large scale database applications). This paper summarizes some of the lessons learned.
 
Keywords: Unified modeling language, reusable object oriented design.
 

 
TRANSFORMATION FRAMEWORK FOR LEGACY SOFTWARE MIGRATION
N. I. Yusop, K. R. Ku-Mahamud and A.T. Othman
School of Information Technology, Universiti Utara Malaysia, 06010 Sintok, Kedah, Malaysia
 
 
ABSTRACT FULL TEXT
A legacy system is a system that is built using old technologies, but it is still beneficial to an organization. However, today it is facing a great challenge to meet the demands of current applications. Organizations are coming under great pressure to decide on the fate of the legacy system as they try to cope with the changing scenario. Among the alternatives offered are: discarding the legacy system and replacing it with a totally new system, allowing the system to deteriorate until the organization is out of business, redeveloping the system, or reengineering the system. Reengineering seems to offer the best solution to the challenge. This paper focuses on transformation, which is one of the reengineering technologies for migrating a legacy system toward an evolvable system. A few legacy system software transformation approaches are discussed. A framework for legacy software migration that combines the strengths of each of the described approaches is proposed.
 
Keywords: Program transformation, software migration, legacy system.

Universiti Utara Malaysia Press
 Universiti Utara Malaysia, 06010 UUM Sintok
Kedah Darul Aman, MALAYSIA
Phone: +604-928 4816, Fax : +604-928 4792

All Right Reserved. Copyright © 2010, Universiti Utara Malaysia Press