)
tracks, where d was the density of the
channel routing problem. The result of this paper is better than this, if d
is very close to its upper bound, namely to l.
In fact, this is rarely the case.AGAPE: parallel
genetic algorithm programming environment developed for ape100/quadrics
A.
Sternieri, P. Anelli, S. Stramaglia, U. Emiliani
Abstract.
AGAPE, an environment for the
implementation of parallel evolutionary algorithms on the APE100/Quadrics
architecture, is presented. AGAPE is a flexible tool for users to solve
numerical optimization problems; in addition researchers can use AGAPE for the
study of new genetic operators, selection functions and migration strategies. A
version of AGAPE dedicated to the design of neural networks of arbitrary kind
and activation function is described and an original comparison among different
evolutionary algorithms to learn neural networks is reported.
The simulation results show that the coding scheme adopted by AGAPE leads
to the best learning rate on Rumelhart's problems. The paper also includes two
reviews on parallel genetic algorithms and on evolution programs for the
optimization of neural networks.
Incremental
view materialization in deductive databases
Wang-Chan
Wong, L.F. Bic
Abstract.
This paper presents a unifying
approach to processing of (recursive) queries and updates in a deductive
database. To improve
query performance, a combined top-down and bottom-up evaluation method is used
to compile rules into iterative programs that contain relational algebra
operators. This method is based
on the lemma resolution that retains previous results to guarantee
termination.
Due
to locality in database processing
(i.e. repetitive user query patterns), it is desirable to materialize frequently
used queries against views of the
database. Unfortunately, if updates are allowed, maintaining materialized views
tables becomes a major problem. We propose to materialize views incrementally,
as queries are being answered. Hence views in our approach are only partially materialized. For such views, we design algorithms to
perform updates only when the
underlying view tables are actually affected.
We
compare our approach to two well-known methods for dealing with views: total
materialization and query-modification.
The first method materializes the entire view when it is defined while the
second recomputes the view on the fly without maintaining any physical view
tables. We demonstrate that our approach is a compromise between these two
methods by determining the conditions under which it performs better.
Improving
performances of the genetic algorithm
by caching
J.
Kratica
Abstract.
In this paper we optimize
run-time performance of the genetic algorithm by caching. We are caching the
genetic algorithm procedure for evaluation of an objective function. Least
Recently Used (LRU) caching strategy is used, that is simple but effective. This
approach is good for problems that have a relatively small length of item
string, and a large evaluation time of objective function. We present results of
the caching to genetic algorithm for solving one such problem - the simple plant
location problem (SPLP).
a
multimedia documentation environment supports
well-engineered software development
and maintenance
T.K.
Shih, L.R. Chow. H.-C. Keh, Y.C. Lin
Abstract.
Program documentation is very
important to software design, coding, testing and maintenance. A well-designed
documentation should reduce the development time and cost, and make the software
more reliable and easier to maintain. But current program documentation has a
number of drawbacks, such as the incompleteness, inconsistency, traceability
problems, no quantitative methods to measure the quality,
and unfriendly
to read and write. These drawbacks cause naïve or maintenance
programmers unwilling to read documents, thus hard to understand the program. In
this paper, we propose a system entitled DocMetrics, which provides four tools
to assist program documentation.
An editor is used to facilitate programmers using multimedia to annotate
their program in a different way. A composer constructs the program into a tree,
integrates the documents, and measures the completeness of documentation. A
browser allows programmers to traverse a program in a hypertext-like way. A
navigator helps the project manager to produce a guided tour of the program that
can lead naïve or maintenance programmers to traverse and understand the
program. The learning status, feedback, and the quality of documentation can be
analysed quantitatively.
Width-reduction
for a linear time channel routing algorithm
M.-A.
Lengyel
Abstract.
This paper presents a new
upper bound for channel routing of multiterminal nets on two layers. The result,
which is essentially the improving of Recski and Strzyzewski's algorithm [2],
works in linear time and uses width at most 4l/3,
where l is the length of the channel. (The aforementioned algorithm used
width at most 3l/2.)
In
1994, Gao and Kaufmann [1] presented a new algorithm for channel routing of
multiterminal nets on two layers, which required 3d/2
+ O (
)
tracks, where d was the density of the
channel routing problem. The result of this paper is better than this, if d
is very close to its upper bound, namely to l.
In fact, this is rarely the case.