Service feature modeling: modeling and participatory ranking of service design alternatives

The design of software-intensive service systems involves and affects numerous stakeholders including software engineers, legal and business experts as well as a potentially large number of consumers. In consequence, the challenge arises to adequately represent the interests of these groups with respect to service design decisions. Specifically, shared service design artifacts and participatory methods for influencing their development in consensus are required, which are not yet state of the art in software service engineering. To this end, we present service feature modeling. Using a modeling notation based on feature-oriented analysis, our approach can represent and interrelate diverse service design concerns and capture their potential combinations as service design alternatives. We further present a method that allows stakeholders to rank service design alternatives based on their preferences. The ranking can support service engineers in selecting viable alternatives for implementation. To exploit this potential, we have implemented a toolkit to enable both modeling and participative ranking of service design alternatives. It has been used to apply service feature modeling in the context of public service design and evaluate the approach in this context.     

Capacitive depression sensor for microfluidic pneumatic networks

This paper reports a capacitive depression sensor intended to make measurements in-line in microfluidics pneumatic networks which works with negative pressures in the order of hundreds of millibars. The principle of operation of the sensor is neither thermal nor magnetic, but purely mechanical, and does not need a servomechanism. The proposed device is composed by two circular membranes linked by a column. This structure presents a particular behavior when exerting negative pressures between the membranes, named negative behavior. The behavior is confirmed prior to fabrication using numerical simulations. The materials used are the negative epoxy photoresist SU-8, printed circuit board (PCB) and sputtered gold. The device has been experimentally tested in the laboratory with successful results, providing an increment of capacitance from 1.7 to 3 pF for applied depressions from 0 to 500 mbar and presenting good agreement with the electromechanical FEM simulations.     

Automatic mapping of parallel applications on multicore architectures using the Servet benchmark suite

Servet is a suite of benchmarks focused on detecting a set of parameters with high influence on the overall performance of multicore systems. These parameters can be used for autotuning codes to increase their performance on multicore clusters. Although Servet has been proved to detect accurately cache hierarchies, bandwidths and bottlenecks in memory accesses, as well as the communication overhead among cores, up to now the impact of the use of this information on application performance optimization has not been assessed. This paper presents a novel algorithm that automatically uses Servet for mapping parallel applications on multicore systems and analyzes its impact on three testbeds using three different parallel programming models: message-passing, shared memory and partitioned global address space (PGAS). Our results show that a suitable mapping policy based on the data provided by this tool can significantly improve the performance of parallel applications without source code modification.     

Real-time mechanoluminescence sensing of the amplitude and duration of impact stress

The present paper reports the real-time sensing of the amplitude and duration of impact stress using mechanoluminescence (ML) of the films such as ZnS:Mn and SrAl₂O₄:Eu. After the impact of a small ball from a low height onto the film, initially the elastico mechanoluminescence (EML) intensity increases with time, attains a peak value and then it decreases with time, initially at a fast rate and later on at a slow rate. The fast decay time of the EML intensity is related to the rate constant for the rise of impact stress and the slow decay time of EML is equal to the lifetime of electrons in the shallow traps lying in the normal piezoelectric region of the crystals, which get filled during the detrapping of thermally stable traps at the time of the increase of pressure. Both the peaks of EML intensity and total EML intensity increase linearly with the height through which the ball is dropped onto the films. The EML spectra are similar to the corresponding photoluminescence and electroluminescence spectra. On the basis of the localized piezoelectrically induced electron detrapping model, expressions are derived for different parameters of the impact stress-induced EML of the films, whereby a good agreement is found between the experimental and theoretical results. As the EML intensity depends on the impact stress, the impact stress can be sensed by measuring the EML intensity. Furthermore, the duration of stress is related to the time t_m corresponding to the peak of the EML intensity versus time curve; hence, the pulse duration of the impact stress can be monitored by measuring the value of time t_m.     

Analysis of complexity indices for classification problems: Cancer gene expression data

Currently, cancer diagnosis at a molecular level has been made possible through the analysis of gene expression data. More specifically, one usually uses machine learning (ML) techniques to build, from cancer gene expression data, automatic diagnosis models (classifiers). Cancer gene expression data often present some characteristics that can have a negative impact in the generalization ability of the classifiers generated. Some of these properties are data sparsity and an unbalanced class distribution. We investigate the results of a set of indices able to extract the intrinsic complexity information from the data. Such measures can be used to analyze, among other things, which particular characteristics of cancer gene expression data mostly impact the prediction ability of support vector machine classifiers. In this context, we also show that, by applying a proper feature selection procedure to the data, one can reduce the influence of those characteristics in the error rates of the classifiers induced.     

Design of decentralized controllers for MIMO processes

In this paper, the design of decentralized PID controller for interactive and delay time processes is proposed. Direct controller synthesis approach is used to obtain parameters of the multi-loop controllers. A detuning factor for each loop is specified based on the closed-loop time constant. Then appropriate controller settings are determined using the Maclaurin series expansion and the model of the process. The success of the proposed algorithm is based on the detuning parameter thus the selection of the desired closed-loop time constant is an important factor. The proposed method has no limitations regarding systems weak or strong interactions, multiple delays, and its operating behaviors such as monotonic or oscillatory. The proposed algorithm is simpler and improved performance of various processes are achieved as compared to some of the prevalent techniques. Simulation examples demonstrate the usefulness of the proposed method.