CFP - IJAST Special issue on Software Component Maintenance and Reuse

[ia...@acm.org]

International Journal of Applied Software Technology (IJAST)

Call For Papers: Special Issue on Software Component Maintenance and
Reuse Through Evolutionary Programming Technology

Guest Editor: David C. Rine

IJAST is organizing a special issue on Software Component Maintenance
and Reuse through Evolutionary Programming Technology to appear in
August 1997. The maintenance of software code components comprises a
major part of their lifecycle. Traditional maintenance approaches for
such components are heavily dependent upon direct programmer
involvement in the adaptations and corrections necessary to improve
their operational quality. Prior to maintenance, developers, on the
other hand, try to make software code components reusable.
Traditionally, reusable software code components have been divided
into the following general classes: components which can be reused as
is without any modifications, components which can be reused with
minor changes, components which can be reused after major changes,
and components which cannot be reused for the given product or
products. Once again, making code components reusable for a given
product, given that modifications are necessary, involves direct
programmer involvement in making these modifications to the code.
Because of the effort, human labor intensiveness and investment
involved in using the present paradigm of software code component
maintenance and reuse, a new and more effective or efficient paradigm
is called for.

Hence, the new software code component maintenance and reuse paradigm
shift assumes the following framework:

(1) The modules will be described using fuzzy variables to model the
specification layer and using fuzzy membership functions to
model the representation layer.

(2) The modules will be maintained and reused by training, i.e.
changing the fuzzy membership function data values, then using
evolutionary programming technologies such as neural networks,
genetic algorithms and Monte Carlo methods.

(3) Thresholds will be established so that the improved
effectiveness of the modules can be measured and evaluated
during training.

(4) With each module there will be stored the set of scenarios (test
cases) that were used in the initial training of the module, and
when the application world scenarios change the new scenarios
will be merged with the initial scenarios followed by the re-
training of the modules using these new scenarios.

(5) The reusability paradigm will now be comprised of a repository
of these modules and their scenarios. The scenario driver will
take the role of 'tester', and the evolutionary programs will
take the role of the 'debugger'.

The first set of applications we perceive that will be developed
using this new paradigm will be those in the 'control' domain where
the fuzzy approach has already met with considerable success in the
products marketplace, e.g. cameras, automobiles, trains, chemical
processes, equipment monitors, etc. We perceive this approach as
becoming a real alternative to software code component maintenance
and reuse in the domains of transportation, engineering processes,
equipment monitoring, and many other areas which can be thought of as
control problems. This approach can even be thought of as being of
use in the maintenance of command and control systems.

TOPICS OF INTEREST UNDER THE ABOVE THEME

artificial neural network paradigms genetic algorithm paradigms
learning algorithms and training fuzzy logic and systems
evolutionary programming soft computing
machine learning adaptive control
model-bases in the control domain diagnosis
process monitoring non-linear systems modeling
evolutionary software maintenance evolutionary software reuse
maintenance and reuse of software code components
control and manufacturing applications

Topics on the applications of evolutionary programming to problems in
software code and code component maintenance and reuse are especially
desired.

Format of submission: Papers should not exceed 30 double spaced pages
(approx. 5000 words). Papers should not have been previously
published, not currently submitted elsewhere for publication. Papers
should include a title page containing title, authors' names and
affiliations, postal and e-mail addresses, telephone numbers and Fax
numbers. Papers should include a 300 word abstract and 5-10 keywords.
Authors should send six copies of their manuscript to David Rine at
the address below.

Schedule: Paper should be submitted by Dec. 14, 1996. Notification of
acceptance will be given by April 15, 1997. Camera ready copy will be
due June 1, 1997.

ADDRESS: David C. Rine, Professor, Department of Computer Science,
Software Systems Engineering Group, School of Information
Technology and Engineering, George Mason University,
Fairfax, VA 22030-4444, USA
PH 703-993-1530 or 1546, FX 703-993-3729, dr...@cne.gmu.edu.