CitationDownload as .RIS
Emerald Group Publishing Limited
Article Type: Editorial From: Kybernetes, Volume 43, Issue 7
There are ten papers in this issue of the journal. They are all highly applied, examining problems including e-commerce, manufacturing, robotics, conflict analysis, economic forecasting and credit scoring. They use techniques ranging from genetic algorithms, fuzzy set theory, image processing, mathematical modelling and neural networks. Once again these papers, which come from a wide variety of countries and cultures, show the breadth of the fields of cybernetics, systems and management science.
In these papers:
Chen, Chiu, Huang and Yeh discuss the design of web sites for online shopping, and the difficulty of optimising both product image allocation and inventory control. They developed a model of the relationship between independent variables, and explore its optimisation using genetic algorithms. They illustrate their algorithms with specific examples.
Liu also discusses issues around e-commerce: the best ways to provide advertising based on check-ins within social media sites .The author draws on media richness and achievement motivation theory to develop a set of hypotheses concerning the influence of social cues on advertising effectiveness. These hypotheses were tested in an experiment using Facebook. The author discusses the results and examines their managerial implications.
Ma and Zhang write from within the field of human-robot interaction, a growing area given the rise of service robots. They present a new method for removing the background from a robot's image of the world, enabling it to focus on more important foreground images. Their method is based on image parameters, representing the changing features of the image. They describe their improved method in some detail and describe experimental results of its use.
Samuel and Venkumar present a problem within manufacturing: the management of production logistics, and in particular the scheduling of a number of machines within a flow shop. They discuss a new algorithm for such scheduled, based on a hybrid form of simulated annealing. They analyse their algorithm with respect to three benchmark problems, concluding that their algorithm is more effective than others.
Liu, Lin and Liu look at conflict analysis, and build upon a classic agent-based conflict theory by Pawlak. From this theory, they build a model using intuitionistic fuzzy sets, which they are then able to apply to an analysis of a real international conflict. They demonstrate the flexibility and broad range of conflicts able to be modelled using their techniques.
Pan examines the forecasting of gold and oil prices, a key issue in advanced economies. The author uses techniques from the field of evolutionary computing to produce more accurate models, drawing upon a modified fruit fly optimisation algorithm that is built into a generalised neural network. The paper describes the development of this algorithm, its application and an evaluation of its use.
Xie and Xin examine a problem of multi-attribute decision making: the selection of suppliers under conditions of uncertain information. They address this problem using a decision-making model derived from grey systems theory. Although their presentation is largely mathematical, it arises from a practical problem and presents a clear algorithm for addressing that problem.
Hadi-Vencheh, Ghelej Beigi and Gholami look at issues of resource allocation and target setting in management science, using the approach of data envelopment analysis (DEA). They use a fuzzy DEA model to examine the introduction of new sub-decision-making units into a network. Again their presentation is largely mathematical, but applied to a practical problem.
Wu, Hu and Huang use machine-learning techniques to enhance the capacity of agencies to conduct credit rating. They examine a variety of different possible machine learning techniques, arguing for the benefits of a two-stage classification technique called Bagging. They apply this technique to a data set describing a large number of Taiwanese companies across key variables, and evaluate it against other possible techniques to demonstrate its effectiveness.
Tsai and Hung look at the related issue of credit scoring. They argue that neural networks are superior to statistical approaches for effective credit scoring. Using data sets drawn from Australia, Germany and Japan, they look at the most effective neural network techniques for credit scoring, in particular advocating the use of hybrid and ensemble neural networks.
We hope you enjoy this issue.
Magnus Ramage, David Chapman and Chris Bissell