A hybrid genetic model for the prediction of corporate failure

This study examines the potential of a neural network (NN) model, whose inputs and structure are automatically selected by means of a genetic algorithm (GA), for the prediction of corporate failure using information drawn from financial statements. The results of this model are compared with those of a linear discriminant analysis (LDA) model. Data from a matched sample of 178 publicly quoted, failed and non-failed, US firms, drawn from the period 1991 to 2000 is used to train and test the models. The best evolved neural network correctly classified 86.7 (76.6)% of the firms in the training set, one (three) year(s) prior to failure, and 80.7 (66.0)% in the out-of-sample validation set. The LDA model correctly categorised 81.7 (75.0)% and 76.0 (64.7)% respectively. The results provide support for a hypothesis that corporate failure can be anticipated, and that a hybrid GA/NN model can outperform an LDA model in this domain.MSC codes: 62M45, 68W10, 90B50, 91C20     

Comparison of ANN and DoE for the prediction of laser-machined micro-channel dimensions

This paper presents four models developed for the prediction of the width and depth dimensions of CO₂ laser-formed micro-channels in glass. A 3³ statistical design of experiments (DoE) model was built and conducted with the power (P), pulse repetition frequency (PRF), and traverse speed (U) of the laser machine as the selected parameters for investigation. Three feed-forward, back-propagation artificial neural networks (ANNs) models were also generated. These ANN models were varied to investigate the influence of variations in the number and the selection of training data. Model A was constructed with 24 data randomly selected from the experimental results, leaving three data points for model testing; Model B was constructed with the eight corner points of the experimental data space, and seven other randomly selected data, leaving 12 data points for testing; and Model C was constructed with 15 randomly selected data leaving 12 data points for testing. These models were developed separately for both micro-channel width and depth prediction. These ANN models were constructed in LabVIEW coding. The performance of these ANN models and the DoE model were compared. When compared with the actual results two of the ANN models showed greater average percentage error than the DoE model. The other ANN model showed an improved predictive capability that was approximately twice as good as that provided from the DoE model.     

A ribose sugar conformational switch in the LTR-retrotransposon Ty3 polypurine tract-containing RNA/DNA hybrid

To probe structural features of a polypurine tract (PPT) that mediate its specific recognition and processing, a model 20 bp RNA/DNA hybrid duplex, which includes the full PPT sequence of the Saccharomyces cerevisiae LTR-retrotransposon Ty3, has been investigated using solution NMR spectroscopy. While homonuclear NOESY and DQF-COSY analyses indicate that this PPT-containing RNA/DNA hybrid adopts an overall A-form-like helical geometry, an unexpected sugar pucker switch has been detected for the ribose at position +1, relative to the cleavage site, on the RNA strand. A model of the conformational changes induced by the A- to B-type sugar pucker switch shows a significant change in the backbone trajectory of the RNA strand, which alters the presentation of backbone phosphate and 2' hydroxyl groups 3' of this residue. This observation implies that part of the mechanism governing RNase H fidelity may be through distortion of the RNA/DNA helix one base ahead of the scissile bond.     

On Markovian approximations for virtual-build-to-order systems

The Virtual Build-to-Order (VBTO) approach strives to allow a producer to fulfil customers with the specific product variants they seek more efficiently than a conventional order fulfilment system. It does so by opening the planning pipeline. Here the feasibility of modelling the VBTO system as a Markov process is investigated. Two system configurations are considered—a random pipeline feed policy that assumes only knowledge of the overall demand pattern and an informed policy that ensures a mix of different variants in the system. First-order Markov models, which assume stationarity requirements are satisfied, are developed for small VBTO systems. The model for the informed feed policy has excellent agreement with simulation results and confirms the superiority of this policy over the random policy. The model for the random policy is more accurate at high variety than at low variety levels. Accuracy is improved with a second-order Markov model. Although impractical for modelling large scale VBTO systems for either configuration, the Markov approach is valuable in providing insights, theoretical foundations and validation for simulation models. It aids the interpretation of observations from simulations of large scale systems and explains the mechanism by which an unrepresentative stock mix develops over time for the random policy.     

Process mapping of laser surface modification of AISI 316L stainless steel for biomedical applications

A 1.5-kW CO₂ laser in pulsed mode at 3 kHz was used to investigate the effects of varied laser process parameters and resulting morphology of AISI 316L stainless steel. Irradiance and residence time were varied between 7.9 to 23.6 MW/cm² and 50 to 167 μs, respectively. A strong correlation between irradiance, residence time, depth of processing and roughness of processed steel was established. The high depth of altered microstructure and increased roughness were linked to higher levels of both irradiance and residence times. Energy fluence and surface temperature models were used to predict levels of melting occurring on the surface through the analysis of roughness and depth of the region processed. Microstructural images captured by the SEM revealed significant grain structure changes at higher irradiances, but due to increased residence times, limited to the laser in use, the hardness values were not improved.     

Atomic diffusion in laser surface modified AISI H13 steel

This paper presents a laser surface modification process of AISI H13 steel using 0.09 and 0.4 mm of laser spot sizes with an aim to increase surface hardness and investigate elements diffusion in laser modified surface. A Rofin DC-015 diffusion-cooled CO₂ slab laser was used to process AISI H13 steel samples. Samples of 10 mm diameter were sectioned to 100 mm length in order to process a predefined circumferential area. The parameters selected for examination were laser peak power, pulse repetition frequency (PRF), and overlap percentage. The hardness properties were tested at 981 mN force. Metallographic study and energy dispersive X-ray spectroscopy (EDXS) were performed to observe presence of elements and their distribution in the sample surface. Maximum hardness achieved in the modified surface was 1017 HV_0.1. Change of elements composition in the modified layer region was detected in the laser modified samples. Diffusion possibly occurred for C, Cr, Cu, Ni, and S elements. The potential found for increase in surface hardness represents an important method to sustain tooling life. The EDXS findings signify understanding of processing parameters effect on the modified surface composition.