Abstract: The importance of supply chain and logistics
management has been widely recognised. Effective management of
the supply chain can reduce costs and lead times and improve
responsiveness to changing customer demands. This paper proposes a
multi-matrix real-coded Generic Algorithm (MRGA) based
optimisation tool that minimises total costs associated within supply
chain logistics. According to finite capacity constraints of all parties
within the chain, Genetic Algorithm (GA) often produces infeasible
chromosomes during initialisation and evolution processes. In the
proposed algorithm, chromosome initialisation procedure, crossover
and mutation operations that always guarantee feasible solutions
were embedded. The proposed algorithm was tested using three sizes
of benchmarking dataset of logistic chain network, which are typical
of those faced by most global manufacturing companies. A half
fractional factorial design was carried out to investigate the influence
of alternative crossover and mutation operators by varying GA
parameters. The analysis of experimental results suggested that the
quality of solutions obtained is sensitive to the ways in which the
genetic parameters and operators are set.
Abstract: Ensemble learning algorithms such as AdaBoost and
Bagging have been in active research and shown improvements in
classification results for several benchmarking data sets with mainly
decision trees as their base classifiers. In this paper we experiment to
apply these Meta learning techniques with classifiers such as random
forests, neural networks and support vector machines. The data sets
are from MAGIC, a Cherenkov telescope experiment. The task is to
classify gamma signals from overwhelmingly hadron and muon
signals representing a rare class classification problem. We compare
the individual classifiers with their ensemble counterparts and
discuss the results. WEKA a wonderful tool for machine learning has
been used for making the experiments.
Abstract: Majority of Business Software Systems (BSS)
Development and Enhancement Projects (D&EP) fail to meet criteria
of their effectiveness, what leads to the considerable financial losses.
One of the fundamental reasons for such projects- exceptionally low
success rate are improperly derived estimates for their costs and time.
In the case of BSS D&EP these attributes are determined by the work
effort, meanwhile reliable and objective effort estimation still appears
to be a great challenge to the software engineering. Thus this paper is
aimed at presenting the most important synthetic conclusions coming
from the author-s own studies concerning the main factors of
effective BSS D&EP work effort estimation. Thanks to the rational
investment decisions made on the basis of reliable and objective
criteria it is possible to reduce losses caused not only by abandoned
projects but also by large scale of overrunning the time and costs of
BSS D&EP execution.
Abstract: Leo Breimans Random Forests (RF) is a recent
development in tree based classifiers and quickly proven to be one of
the most important algorithms in the machine learning literature. It
has shown robust and improved results of classifications on standard
data sets. Ensemble learning algorithms such as AdaBoost and
Bagging have been in active research and shown improvements in
classification results for several benchmarking data sets with mainly
decision trees as their base classifiers. In this paper we experiment to
apply these Meta learning techniques to the random forests. We
experiment the working of the ensembles of random forests on the
standard data sets available in UCI data sets. We compare the
original random forest algorithm with their ensemble counterparts
and discuss the results.