International Joint Conference SOCO’16-CISIS’16-ICEUTE’16

Objective: Predicting Emergency Department (ED) readmissions is of great importance since it helps identifying patients requiring further post-discharge attention as well as reducing healthcare costs. It is becoming standard procedure to evaluate the risk of ED readmission within 30 days after discharge. Methods. Our dataset is stratified into four groups according to the Kaiser Permanente Risk Stratification Model. We deal with imbalanced data using different approaches for resampling. Feature selection is also addressed by a wrapper method which evaluates feature set importance by the performance of various classifiers trained on them. Results. We trained a model for each scenario and subpopulation, namely case management (CM), heart failure (HF), chronic obstructive pulmonary disease (COPD) and diabetes mellitus (DM). Using the full dataset we found that the best sensitivity is achieved by SVM using over-sampling methods (40.62 % sensitivity, 78.71 % specificity and 71.94 accuracy). Conclusions. Imbalance correction techniques allow to achieve better sensitivity performance, however the dataset has not enough positive cases, hindering the achievement of better prediction ability. The arbitrary definition of a threshold-based discretization for measurements which are inherently is an important drawback for the exploitation of the data, therefore a regression approach is considered as future work.

This paper employs well-known techniques as Support Vector Machines and Neural Networks in order to classify images of boar sperm cells. Acrosome integrity gives information about if a sperm cell is able to fertilize an oocyte. If the acrosome is intact, the fertilization is possible. Otherwise, if a sperm cell has already reacted and has lost its acrosome or even if it is going through the capacitation process, such sperm cell has lost its capability to fertilize. Using a set of descriptors already proposed to describe the acrosome state of a boar sperm cell image, two different classifiers are considered. Results show the classification accuracy improves previous results.

Wearable devices are currently used in researches related with the detection of human activities and the anamnesis of illnesses. Recent studies focused on the detection of simulated epileptic seizures have found that Fuzzy Rule Base Classifiers (FRBC) can be learnt with Ant Colony Systems (ACS) to efficiently deal with this problem. However, the computational requirements for obtaining these models is relatively high, which suggests that an alternative for reducing the learning cost would be rather interesting. Therefore, this study focuses on reducing the complexity of the model by using a discretization technique, more specifically, the discretization proposed in the SAX Time Series (TS) representation.Therefore, the very simple discretization method based on the probability distribution of the values in the domain is used together with the AntMiner+ and a Pittsburg FRBC learning algorithm using ACS. The proposal have been tested with a realistic data set gathered with participants following a very strict protocol for simulating epileptic seizures, each participant using a wearable device including tri-axial accelerometers placed on the dominant wrist.The experimentation shows that the discretization method has clearly improved previous published results. In the case of Pittsburg learning, the generalization capabilities of the models have been greatly enhanced, while the models learned with this partitioning and the AntMiner+ have outperformed all the models in the comparison. These results represent a promising starting point for the detection of epileptic seizures and will be tested with patients in their own environment: it is expected to start gathering this data during the last quarter of this year.

This paper presents a real solution for detecting abnormal situations at home environments, mainly oriented to living alone and elderly people. The aim of the work described in this paper is, first, to reduce the raw data about the situation of the elder at home, tracking only the relevant signals, and second, to predict the regular situation of the person at home, checking if its situation is normal or abnormal. The challenge in this work is to transform the real word complexity of the user patterns using only “lazy” sensor data (position sensors) in a real scenario over several homes. We impose two restrictions to the system (lack of “a priori” information about the behavior of the elderly and the absence of historic database) because the aim of this system is to build an automatic environment and study the minimal historical data to achieve an accurate predictive model, in order to generate a commercial produtc working fully few weeks after the installation.

Neuroimaging data provides a valuable tool to assist the diagnosis of neurodegenerative disorders such as Alzheimer’s disease (AD) and Parkinson’s disease (PD). During last years many research efforts have focused on the development of computer systems that automatically analyze neuroimaging data and allow improving the diagnosis of those diseases. This field has benefited from modern machine learning techniques, which provide a higher generalization ability, however the high dimensionality of the data is still a challenge and there is room for improvement. In this work we demonstrate a computer system based on Principal Component Analysis and TensorFlow, the machine learning library recently released by Google. The proposed system is able to successfully separate AD or PD patients from healthy subjects, as well as distinguishing between PD and other parkinsonian syndromes. The obtained results suggest that TensorFlow is a suitable environment to classify neuroimaging data and can help to improve the diagnosis of AD and Parkinsonism.

A wide range of industrial fields utilize cyclone separators and so, evaluating their performance according to different materials and varying operating conditions could contribute useful information and could also save these industries significant amounts of capital. This study models cyclone performance using linear regression techniques and low errors were obtained in comparison with the values obtained from real experiments. Linear regression and generalized linear regression techniques, simple and enhanced with Gradient Boosting techniques, were used to create linear models with low errors of approximately 0.83 % in cyclone performance.

This study presents the application of clustering techniques to a real-life problem of studying the air quality of the Castilla y León region in Spain. The goal of this work is to analyze the level of air pollution in eight points of this Spanish region between years 2008 and 2015. The analyzed data were provided by eight acquisition stations from the regional network of air quality. The main pollutants recorded at these stations are analyzed in order to study the characterization of such stations, according to a zoning process, and their time evolution. Four cluster evaluation and a clustering technique, with the main distance measures, have been applied to the dataset under analysis.

This paper proposes the use of intelligent wristbands for the automatic detection of emotional states in order to develop an application which allows to extract, analyze, represent and manage the social emotion of a group of entities. Nowadays, the detection of the joined emotion of an heterogeneous group of people is still an open issue. Most of the existing approaches are centered in the emotion detection and management of a single entity. Concretely, the application tries to detect how music can influence in a positive or negative way over individuals’ emotional states. The main goal of the proposed system is to play music that encourages the increase of happiness of the overall patrons.

An impulse noise detection scheme employing machine learning (ML) algorithm in Orthogonal Frequency Division Multiplexing (OFDM) is investigated. Four powerful ML’s multi-classifiers (ensemble) algorithms (Boosting (Bos), Bagging (Bag), Stacking (Stack) and Random Forest (RF)) were used at the receiver side of the OFDM system to detect if the received noisy signal contained impulse noise or not. The ML’s ensembles were trained with the Middleton Class A noise model which was the noise model used in the OFDM system. In terms of prediction accuracy, the results obtained from the four ML’s Ensembles techniques show that ML can be used to predict impulse noise in communication systems, in particular OFDM.

In the present day, one of the most common activities of everyday life is going to a supermarket or similar retail spaces to buy groceries. Many consumers organizations like The European Consumer Organization [1], advise buyers to prepare a “grocery list” in order to be ready for this activity. The present work proposes a system that helps to develop this activity in several ways: Firstly, it enables the user to create lists with different levels of abstraction: from concrete products to generic ones (or families of products). Secondly, the lists are collaborative and can be shared with other users. Finally, it automatically determines the best store to buy a given product using the proposed optimization algorithm. Furthermore, the optimization algorithm assigns a part of the list to each user balancing the cost that every user has to pay and choosing the cheapest supermarket where they have to buy.

Empirical models are widely used to estimate solar radiation at locations where other more readily available meteorological variables are recorded. Within this group, soft computing techniques are the ones that provide more accurate results as they are able to relate all recorded variables with solar radiation. In this work, a new implementation of Gradient Boosting Machines (GBMs) named XGBoost is used to predict daily global horizontal irradiation at locations where no pyranometer records are available. The study is conducted with data from 38 ground stations in Castilla-La Mancha from 2001 to 2013.Results showed a good generalization capacity of the model, obtaining an average MAE of 1.63 $$\mathrm{MJ/m}^2$$MJ/m2 in stations not used to calibrate the model, and thus outperforming other statistical models found in the literature for Spain. A detailed error analysis was performed to understand the distribution of errors according to the clearness index and level of radiation. Moreover, the contribution of each input was also analyzed.

This paper deals with the control of the output power gas temperature at the Main Heat Exchanger (MHE). The MHE is the basic part of the Flexible Energy cogeneration System (FES) with the combined Brayton - Rankine cycle, which is designed and constructed at Vitkovice Power Engineering JSC. The FES burns solid fuel and generates electrical energy and thermal energy. The standard temperature control at MHE has two goals. The first control task is the protection of the heat transfer surfaces of the MHE against overheating. The second control task is the stabilization of the temperature of power gas at the output of the MHE which is performed, in principle, by the change of the flow rate of air generated by the compressor and by the change of the flow rate of fuel at the inlet of the FES combustion chamber. In this paper, the control of the temperature of the power gas without the overheating of the heat transfer surfaces of the MHE will be described and analyzed.

Cyber-Physical Systems (CPS) is an emergent approach of physical processes, computer and networking, that focuses on the interaction between cyber and physical elements. These systems monitor and control the physical infrastructures, that is why they have a high impact in industrial automation. The implementation and operation of CPS just like the management of the resulting automation infrastructure is of key importance to the industry. The evolution towards Industry 4.0 is mainly based on digital technologies. We present the integration of Industry 4.0 within the textile industry.

This paper is about the joint operation of wind power with thermal power for bidding in day-ahead electricity market. Start-up and variable costs of operation, start-up/shut-down ramp rate limits, and ramp-up limit are modeled for the thermal units. Uncertainty not only due to the electricity market price, but also due to wind power is handled in the context of stochastic mix integer linear programming. The influence of the ratio between the wind power and the thermal power installed capacities on the expected profit is investigated. Comparison between joint and disjoint operations is discussed as a case study.

In this paper authors address the practical problem of designing an empirical model for a commercial photovoltaic (PV) module (Mitsubishi PV-TD1185MF5) placed at the Faculty of Engineering of Vitoria (Basque Country University, Spain) based on artificial neural networks (ANN). This model obtains Ipv from Vpv, and the paper explains how the empirical data have been gathered and discusses the obtained results. The model reached an average accuracy of 0,15 A and a medium correlation value of R = 0,995.

The energy industry is one of the domains that need control and monitoring at several levels. This article presents a SCADA monitoring system targeting a large scale process which is in need of immediate and frequent interventions: the monitoring and control of an electrical installation supplied by a hydro-generator. Citect SCADA was used to design a flexible solution and can be used for small or large hydro-generators. The SCADA solution created will be used for the functioning optimization, transmission and supervising of functioning programs execution.

Green Logistics are gaining importance due to green house gas emissions and its adverse impact on the environment. In this paper, we address the issues with vehicle routing and emissions. This paper reports the emissions that arise with Vehicle Routing Problem with Mid way Halts (VRPMH) and concentrates in finding low cost route for VRPMH using PSO with local exchange. Along with distance minimization, cruise and idling state emissions are reported. Computational experiments are carried out with green vehicle routing problem instances and the results are tabulated. The results project the impact of idling emissions and the need for its possible reduction.

Most computational models of influence spread nowadays are motivated by the need to identify the social actors with maximal influence, in order to achieve high penetration in the market with minimal effort. However, there are little literature on the mechanisms of influence propagation, i.e. computational models of how the social actors change their opinions. There are some works that relate the spatial distribution of the opinions with the mechanism by which an agent changes or maintains its opinions, but they assume a cell model, where agents have fixed spatial locations and neighbors. Here we explore the effect of spatial interaction of the agents, which are free to move in a given space, following attraction dynamics towards agents with similar opinions. The spatial distribution of opinions observed by the agent is used by the agent to decide about opinion changes. We report preliminary results of simulations carried out in Netlogo environment for the first three kinds of systems.

In this paper, we present a new coevolutionary algorithm for workflow scheduling in a dynamically changing environment. Nowadays, there are many efficient algorithms for workflow execution planning, many of which are based on the combination of heuristic and metaheuristic approaches or other forms of hybridization. The coevolutionary genetic algorithm (CGA) offers an extended mechanism for scheduling based on two principal operations: task mapping and resource configuration. While task mapping is a basic function of resource allocation, resource configuration changes the computational environment with the help of the virtualization mechanism. In this paper, we present a strategy for improving the CGA for dynamically changing environments that has a significant impact on the final dynamic CGA execution process.

EXtreme Gradient Boosting (XGBoost) has become one of the most successful techniques in machine learning competitions. It is computationally efficient and scalable, it supports a wide variety of objective functions and it includes different mechanisms to avoid over-fitting and improve accuracy. Having so many tuning parameters, soft computing (SC) is an alternative to search precise and robust models against classical hyper-tuning methods. In this context, we present a preliminary study in which a SC methodology, named GA-PARSIMONY, is used to find accurate and parsimonious XGBoost solutions. The methodology was designed to optimize the search of parsimonious models by feature selection, parameter tuning and model selection. In this work, different experiments are conducted with four complexity metrics in six high dimensional datasets. Although XGBoost performs well with high-dimensional databases, preliminary results indicated that GA-PARSIMONY with feature selection slightly improved the testing error. Therefore, the choice of solutions with fewer inputs, between those with similar cross-validation errors, can help to obtain more robust solutions with better generalization capabilities.

A genetic algorithm, that exploits the K-means principles for dividing objects in groups having high similarity, is proposed. The method evolves a population of chromosomes, each representing a division of objects in a different number of clusters. A group-based crossover, enriched with the one-step K-means operator, and a mutation strategy that reassigns objects to clusters on the base of their distance to the clusters computed so far, allow the approach to determine the best number of groups present in the dataset. The method has been experimented with four different fitness functions on both synthetic and real-world datasets, for which the ground-truth division is known, and compared with the K-means method. Results show that the approach obtains higher values of evaluation indexes than that obtained by the K-means method.

Human intervertebral lumbar disc degeneration is painful and difficult to treat, and is often magnified when the patient is overweight. When the damage is excessive, the disc is replaced by a non-natural or artificial disc. Artificial discs sometimes have the disadvantage of totally different behavior from that of the natural disc. This affects substantially the quality of treated patient’s life. The Finite Element Method (FEM) has been used for years to design an artificial disc, but it involves a high computational cost. This paper proposes a methodology to design a new Artificial Human Intervertebral Lumbar Disc by combining FEM and soft computing techniques. Firstly, a three-dimensional Finite Element (FE) model of a healthy disc was generated and validated experimentally from cadavers by standard tests. Then, an Artificial Human Intervertebral Lumbar Disc FE model with a core of Polycarbonate Polyurethane (PCU) was modeled and parameterized. The healthy and artificial disc FE models were both assembled between lumbar vertebrae L4-L5, giving place to the Functional Spinal Unit (FSU). A Box-Behnken Design of Experiment (DoE) was generated that considers the parameters that define the geometry of the proposed artificial disc FE model and the load derived from the patient’s height and body weight. Artificial Neural Networks (ANNs) and regression trees that are based on heuristic methods and evolutionary algorithms were used for modeling the compression and lateral bending stiffness from the FE simulations of the artificial disc. In this case, ANNs proved to be the models that had the best generalization ability. Finally, the best geometry of the artificial disc proposed when the patient’s height and body weight were considered was achieved by applying Genetic Algorithms (GA) to the ANNs. The difference between the compression and lateral bending stiffness obtained from the healthy and artificial discs did not differ significantly. This indicated that the proposed methodology provides a powerful tool for the design and optimization of an artificial prosthesis.

The paper deals with a machine vision for a recognition and identification of manufactured parts. The inspection line has been constructed. Stepper motors, direct current motors and electromagnetic coils have been used as actuators. Printed circuit boards have been designed and made for a power supplying and signal level conversion. A feedback is acquired by two–position switches. For the practical realization, various construction components have been used, for example buffers for objects prepared for the inspection and inspected objects, manipulation arms and manipulation platform. The main control unit is a compact programmable logical controller, which controls hardware parts during the inspection cycle. The image capturing has been done with a common web camera. Algorithms for the image processing has been programmed in MATLAB Simulink. Data between the control system and the image processing system are exchanged via a mutual communication protocol.

Among current foreground detection algorithms for video sequences, methods based on self-organizing maps are obtaining a greater relevance. In this work we propose a probabilistic self-organising map based model, which uses a uniform distribution to represent the foreground. A suitable set of characteristic pixel features is chosen to train the probabilistic model. Our approach has been compared to some competing methods on a test set of benchmark videos, with favorable results.

The implementation of a reliable vision system for full perception of the human-robot environment is a key issue for the flexible collaborative production industries, especially for the frequently changing applications. The use of such system facilitates the perception and recognition of the human activity, and consequently highly increases the robustness and reactivity of safety strategies in collaborative tasks. This paper presents an implementation of several techniques for workspace monitoring in collaborative human-robot applications. A reliable perception of the overall environment is performed to generate a consistent point cloud which is used for human detection and tracking. Additionally, safety strategies on the robotic system (reduced velocity, emergency stop, ...) are activated when the human-robot distance approaches predefined security thresholds.

In this paper, we explore data collected in a pilot project that used a digital camera and facial recognition to detect foot traffic to a sports store. Using a time series approach, we model daily incoming store traffic under three classes (all faces, female, male) and compare six forecasting approaches, including Holt-Winters (HW), a Support Vector Machine (SVM) and a HW-SVM hybrid that includes other data features (e.g., weather conditions). Several experiments were held, under a robust rolling windows scheme that considers up to one week ahead predictions and two metrics (predictive error and estimated store benefit). Overall, competitive results were achieved by the SVM (all faces), HW (female) and HW-SVM (male) methods, which can potentially lead to valuable gains (e.g., enhancing store marketing or human resource management).

The next generation of adaptive optics (AO) systems require tomographic techniques in order to correct for atmospheric turbulence along lines of sight separated from the guide stars. Multi-object adaptive optics (MOAO) is one such technique. Here we present an improved version of CARMEN, a tomographic reconstructor based on machine learning, using a dedicated neural network framework as Torch. We can observe a significant improvement on the training an execution times of the neural network, thanks to the use of the GPU.

Nowadays, the engineering tools play an important role in medicine, regardless of the area. The present research is focused in anesthesiology, specifically on the behavior of sedated patients. The work shows the Bispectral Index Signal (BIS) modeling of patients undergoing general anesthesia during surgery. With the aim of predicting the patient BIS signal, a model that allows to know its performance from the Electromyogram (EMG) and the propofol infusion rate has been created. The proposal has been achieved by using clustering combined with regression techniques and using a real dataset obtained from patients undergoing general anesthesia. Finally, the created model has been tested also with data from real patients, and the results obtained attested the accuracy of the model.

Stress has a big impact in the current society, being the cause or the incentive of several diseases. Therefore, its detection and monitorization has been the focus of a big number of investigations in the last decades. This work proposes the use of physiological variables such as the electrocardiogram (ECG), the galvanic skin response (GSR) and the respiration (RSP) in order to estimate the level and classify the type of stress. On that purpose, an algorithm based on fuzzy logic has been implemented. This computer-intelligent technique has been combined with a structured processing shaped in state machine. This processing classifies stress in 3 different phases or states: alarm, continued stress and relax. An improved estimation of stress level is obtained at the end, considering the last progresses made by different authors. All this is accompanied by stress classification, which is the novelty compared to other works.

Hand grasp is a complex system that plays an important role in the activities of daily living. Upper-limb neuroprostheses aim at restoring lost reaching and grasping functions on people suffering from neural disorders. However, the dimensionality and complexity of the upper-limb makes the neuroprostheses modeling and control challenging. In this work we present preliminary results for checking the feasibility of using a reinforcement learning (RL) approach for achieving grasp functions with a surface multi-field neuroprosthesis for grasping. Grasps from 20 healthy subjects were recorded to build a reference for the RL system and then two different award strategies were tested on simulations based on neuro-fuzzy models of hemiplegic patients. These first results suggest that RL might be a possible solution for obtaining grasp function by means of multi-field neuroprostheses in the near future.

This paper presents a prediction system based on fuzzy modeling of Japanese candlesticks. The prediction is performed using the pattern recognition methodology and applying a lazy and nonparametric classification technique, k-Nearest Neighbours (k-NN). The Japanese candlestick chart summarizes the trading period of a commodity with only 4 parameters (open, high, low and close). The main idea of the decision system implemented in this article is to predict with accuracy, based on this vague information from previous sessions, the performance of future sessions. Therefore, investors could have valuable information about the next session and set their investment strategies.

The search of emerging patterns pursues the description of a problem through the obtaining of trends in the time, or characterisation of differences between classes or group of variables. This contribution presents an application to a real-world problem related to the photovoltaic technology through the algorithm EvAEP. Specifically, the algorithm is an evolutionary fuzzy system for emerging pattern mining applied to a problem of concentrating photovoltaic technology which is focused on the generation of electricity reducing the associated costs. Emerging patterns have discovered relevant information for the experts when the maximum power is reached for the cells of concentrating photovoltaic.

Outdoor air quality monitoring plays crucial role on preventing environment pollution. The idea of use of unmanned aerial vehicles (UAV) in this area is of great interest cause they provide more flexibility than ground systems. The main focus of this work is to propose alternative, competitive outdoor wireless monitoring system that will allow to collect pollution data, detect and locate leakage places within petrol, gas and refinery stations or in hard to reach places. This system should be lightweight, compact, could be mounted on any UAV, operate in GPS denied environments and should be easily deployed and piloted by operator with minimal risk to his health. This paper presents the system, configured on a commercial UAV AR.Drone, embedding gas sensor to it, where as a ground station stands Robot Operation System. Conducted first stage experiments proved capabilities of our system to operate in real-world conditions and serve as a basis to carry out further research.

Published research on manufacturing processes of ceramic insulators concerns mostly material examinations. Little has been done in the field of assuring proper quality of insulators based on analysis of production data. This is why the paper discusses a new approach to supporting quality control in manufacture of ceramic insulators, based on regression analysis and ANN modeling. The proposed algorithm enables the user to control addition of raw aluminum oxide (and its graining) in order to obtain its desired grain-size composition in the mass and thus to reduce the number of defects to acceptable levels.

The paper presents an optimization task of transportation - production solved with genetic algorithms. For the network of processing plants (factories) and collection centers the cost-optimal transportation plan will be established. Plan is regarding to raw materials to the relevant factories. Task of transportation - production regard to the milk transport and processing will be investigated. It is assumed that the functions defining the costs of processing are polynomials of the second degree. Genetic algorithms, their properties and capabilities in solving computational problems will be described and conclusions will be presented. The program that uses genetic algorithms written in MATLAB will be used to solve an investigated issue.

Literature study reveals the existence of many algorithms for detecting cycles (also known as circuits) in directed graphs. Majority of these algorithms can be classified into two groups: (1) traversal algorithms (e.g. variants of depth-first-search algorithms), and (2) matrix-based algorithms (manipulation of the adjacency matrix and its power series). Adjacency matrix based algorithms are computationally simple and more compact than the traversal algorithms. However, a Petri net, due to its bipartite nature, possesses sparse matrix as adjacency matrix. Hence, the matrix based algorithms become inefficient as the algorithm work through redundant data most of its running time. In this paper, we take a closer look into the structure of the adjacency matrix of a Petri net and then propose an algorithm for detecting existence of cycles; the proposed algorithm is computationally efficient as its works through non-redundant data only, as well as simple, easy to understand and easy to implement.

Cutting and packing problems have important applications to the transportation of cargo. Many algorithms have been proposed for solving the 2D/3D cutting stock problems but most of them consider single objective optimization. The goal of the problem here proposed is to load the boxes that would provide the highest total volume and weight to the container, without exceeding the container limits. These two objectives are conflicting because the volume of a box is usually not proportional to its weight. This work deals with a multi-objective formulation of the 3D Packing Problem (3DPP). We propose to apply multi-objective evolutionary algorithms in order to obtain a set of non-dominated solutions, from which the final users would choose the one to be definitely carried out. For doing an extensive study, it would be necessary to use more problem instances. Instances to deal with the multi-objective 3DPP are non-existent. For this purpose, we have implemented an instance generator.

In the paper, the problem of flow planning in production systems belonging to a class of Cyclic Concurrent Processes Systems is presented. The possibility of using Activity-oriented Petri Nets approach to model the problem as a discrete event system model, and then to perform simulations with the tool known as GPenSIM is shown. For simulation dispatching rules achieved by the analytical method of production order verification based on constraints sequencing methodology and its computer implementation in the system of production orders verification SWZ is used. This paper shows that the results achieved from GPenSIM agrees with that from SWZ. The proposed approach is easy to implement and can be used as an effective production plans verification tool, to identify problems during the implementation of the control rules on production resources in the form of deadlocks, failure to disclose a bottleneck, or problems with production flow synchronization.

The paper presents the conception of algorithm for scheduling of manufacturing systems with consideration of flexible resources and production routes. The proposed algorithm is based on ant colony optimisation (ACO) mechanisms. Although ACO metaheuristics do not guarantee finding optimal solutions, and their performance strongly depends on the intensification and the diversification parameters tuning, they are an interesting alternatives in solving NP hard problems. Their effectiveness and comparison with other methods are presented e.g. in [1, 4, 8]. The discussed search space is defined by the graph of operations planning relationships of the set of orders – the directed AND/OR-type graph describing precedence relations between all operations for scheduling. In the structure of the graph the notation ‘operation on the node’ is used. The presented model supports complex production orders, with hierarchical structures of processes and their execution according to both forward and backward strategies.

The paper presents the concept of multistage sequencing system which main task is to aid the processes of production scheduling in modern assembly systems, mainly in automotive industry, but also in other industry branches, where mixed-model production is conducted. Because of NP-hard character of considered scheduling tasks, it is a potential application area of soft computing graph searching techniques. Described sequencing system is a tool for production scheduling optimizing the process in terms defined by the user (e.g. the number of cycle times exceeded on the assembly line stations as well as the average exceeding of the cycle time). The system structure is based on graph searching techniques, and may be implemented in many problems solving methods, including not only production scheduling but also others, e.g. project scheduling. The developed system model is the basis of its implementation in production planning software, creating the sequences of production orders in few different stages.

One of the most important problems consider with the scheduling process is to ensure the needed level of robustness of obtained schedules. One of possible tools that could be used to realize this objective is the Taboo Search Algorithm (TSA). The Average Slack Method (ASM) enables to obtain the best performance of the job shop system. In the paper is presented analysis of two objectives: to achieve the best compromise basic schedule for four efficiency measures as well as to achieve the best compromise reactive schedule. It was investigated of 15 processes executed on 10 machines. It was shown that ASM enables the obtainment of the best performance of the job shop system.

The behavior of a Material Transportation System (MTS) encompassing movement of various transport modes has to be admissible, i.e. collision- and congestion-free, as to guarantee deadlock-free different flows of concurrently transported goods. Since the material flows following possible machining routes serviced by MTS determine its behavior the following questions occur: what kind of MTS structure can guarantee a given behavior, and what admissible behavior can be reachable in a given MTS structure? These questions are typical for vehicle routing problems which are computationally hard. Their formulation within the framework of mesh-like and fractal-like structures enables, however, to get a significant reduction on the size. Such structures enable to evaluate admissible routings and schedules following flow-paths of material transportation in a polynomial time. Considered in the paper production routes followed by MTS are serviced by operations subsequently executed by AGVs and machine tools. The transport operations performed by AGVs are arranged in a streaming closed-loops network where potential conflicts are resolved by priority dispatching rules assigned to shared resources. The main problem boils down to the searching for sufficient conditions guaranteeing MTS cyclic steady state behavior. Implementation of proposed conditions is illustrated through multiple examples.

This paper investigates one approach to the no-wait flow shop scheduling problem with the objective to improve the robustness of created schedules. One of the fundamental objectives is obtaining an optimal solution for this type of complex, large-sized problems in reasonable computational time. For this purpose was used a new hybrid multi-objective algorithm based on the features of a biological immune system (IS) and bacterial optimization (BO) to find Pareto optimal solutions. It is proposed the hybrid multi-objective immune algorithm (HMOIA II). Computational results suggest that proposed HMOIA II enables the obtainment of stable and robust schedules in case of the disturbance.

Scheduling problems are NP-hard, thus have few alternative methods for obtaining solutions. Genetic algorithms have been used to solve scheduling problems; however, the application of genetic algorithms are too expectant, as the steps involved in a genetic algorithm, especially the reproduction step and the selection step, are often time-consuming and computationally expensive. This is because the newly reproduced chromosomes are often redundant or invalid. This paper proposes a brute-force approach for solving scheduling problems, as an alternative to genetic algorithm; the proposed approach is based on Activity-oriented Petri nets (AOPN) and is computationally simple; in addition, the proposed approach also provides the optimal solution as it scans the whole workspace, whereas genetic algorithm does not guarantee optimal solution.

Imperfections in the form of losses and delays in production are visible especially in SME dedicated for MTO. Improving the operation of systems at the stage of product design and production organization and management becomes great challenge. The paper proposes new, holonic-based approach to manufacturing systems integration. In this context authors presents results of their achievements in the area of technical and organizational production preparation, and production running. Proposed SME’s functional modules (that behave like holons) are generally independent. The concept of the systems integration depending on continuously occurring requirements and conditions involving the creation of a process plan, preparation of schedules and the data acquired from production system is proposed in the paper. As a result of implementation of the presented integration method increases the effectiveness in the integrated areas of decision-making, and moreover, the manufacturing abilities of SME companies.

Attackers have evolved classic code-injection attacks, such as those caused by buffer overflows to sophisticated Turing-complete code-reuse attacks. Control-Flow Integrity (CFI) is a defence mechanism to eliminate control-flow hijacking attacks caused by common memory errors. CFI relies on static analysis for the creation of a program’s control-flow graph (CFG), then at runtime CFI ensures that the program follows the legitimate path. Thereby, when an attacker tries to execute malicious shellcode, CFI detects an unintended path and aborts execution. CFI heavily relies on static analysis for the accurate generation of the control-flow graph, and its security depends on how strictly the CFG is generated and enforced.This paper reviews the CFI schemes proposed over the last ten years and assesses their security guarantees against advanced exploitation techniques.

The number of vehicles around the world has increased a lot during the last decades and will continue to grow in the next years. This factor is closely related to the increase of air pollution, traffic congestion and road non-safety. Due to this, the automotive industry and research entities have put lots of efforts in exploring the potential of the Intelligent Transportation Systems (ITS). These efforts have been focused on trying to add ITS systems to the new cars of the future. The present work describes a proposal for taking advantage of ITS possibilities using the vehicles that are nowadays on our roads. In particular, the described platform is mainly based on smartphones, but also on sensors and servers. The proposal is based on a hybrid scheme that combines an online mode using the Internet access of the smartphone, with an offline mode using wireless technologies such as WiFi Direct and Bluetooth Low Energy to define various ITS applications related with traffic collisions, violations, jams, signs and lights, and parking management. The described system has been developed for the Android platform, producing promising results.

In this paper, we propose a new method of learning for speech signal. This technique is based on the deep learning and the wavelet network theories. The goal of our approach is to construct a deep wavelet network (DWN) using a series of Stacked Wavelet Auto-Encoders. The DWN is devoted to the classification of one class compared to other classes of the dataset. The Mel-Frequency Cepstral Coefficients (MFCC) is chosen to select speech features. Finally, the experimental test is performed on a prepared corpus of Arabic words.

Detection of intruders or unauthorized access to computers has always been critical when dealing with information systems, where security, integrity and privacy are key issues. Although more and more sophisticated and efficient detection strategies are being developed and implemented, both hardware and software, there is still the necessity of improving them to completely eradicate illegitimate access. The purpose of this paper is to show how soft computing techniques can be used to identify unauthorized access to computers. Advanced data analysis is first applied to obtain a qualitative approach to the data. Decision tree are used to obtain users’ behavior patterns. Neural networks are then chosen as classifiers to identify intrusion detection. The result obtained applying this combination of intelligent techniques on real data is encouraging.

Even if combined with other techniques, passwords are still the main way of authentication in many services and systems. Attackers can usually test many passwords very quickly when using standard hash functions, so specific password hashing algorithms have been designed to slow down brute force attacks.Spritz is a sponge-based stream cipher intended to be a drop-in replacement for RC4. It is more secure, more complex and more versatile than RC4. Since it is based on a sponge function, it can be employed for other applications like password hashing.In this paper we build upon Spritz to construct a password hashing algorithm and study its performance and suitability.

During the last decade, the flow of 3D objects is increasingly used everywhere. This wide range of applications and the necessity to exchange 3D meshes via internet raise major security problems. As a solution, we propose a blind watermarking algorithm for 3D multi-resolution meshes ensuring a good compromise between invisibility, insertion rate and robustness while minimizing the amount of memory used during the execution of our algorithm. To this end, spiral scanning method is applied. It decomposes the mesh into GOTs (a Group Of Triangles). At each time, only one GOT will be loaded into memory to be watermarked. It undergoes a wavelet transform, a modulation then embedding data. Once finished, the memory will be released to upload the next GOT. This process is stopped when the entire mesh is watermarked. Experimental tests showed that the quality of watermarked meshes is kept despite the high insertion rate used and that memory consumption is very reduced (until 24 % of memory reduction). As for the robustness, our algorithm overcomes the most popular attacks in particular compression. A comparison with literature showed that our algorithm gives better results than those recently published.

Security researchers have been focusing on developing mitigation and protection mechanisms against code-injection and code-reuse attacks. Modern defences focus on protecting the legitimate control-flow of a program, nevertheless they cannot withstand a more subtle type of attack, non-control-data attacks, since they follow the legitimate control flow, and thus leave no trace. Data-Flow Integrity (DFI) is a defence mechanism which aims to protect programs against non-control-data attacks. DFI uses static analysis to compute the data-flow graph of a program, and then, enforce at runtime that the data-flow of the program follows the legitimate path; otherwise the execution is aborted.In this paper, we review the state of the techniques to generate non-control-data attacks and present the state of DFI methods.

The robustness, resilience and security of supply chain transportation is an active research topic, as it directly determines the overall supply chain resilience and security. In this paper, we propose a theoretical model for the transportation problem within a two-stage supply chain network with security constraints called the Secure Supply Chain Network (SSCN). The SSCN contains a manufacturer, directly connected to several distribution centres DC, which are directly connected to one or more customers C. Each direct link between any two elements of Secure Supply Chain Network is allocated a transportation cost. Within the proposed model, the manufacturer produces a single product type; each distribution centre has a fixed capacity and a security rank. The overall objective of the Secure Supply Chain Network is 100 % customer satisfaction whilst fully satisfying the security constraints and minimizing the overall transportation costs. A heuristic solving technique is proposed and discussed.

In this paper we propose modifications to our machine-learning web-layer anomaly detection system that adapts HTTP content mechanism. Particularly we introduce more effective packet segmentation mechanism, adapt AdaBoost classifier, and present results on more challenging dataset. In this paper we also compared our approach with other techniques and reported the results of our experiments.

Cluster forests is a novel approach for ensemble clustering based on the aggregation of partial K-means clustering trees. Cluster forests was inspired from random forests algorithm. Cluster forests gives better results than other popular clustering algorithms on most standard benchmarks. In this paper, we propose an improved version of cluster forests using fuzzy C-means clustering. Results shows that the proposed Fuzzy Cluster Forests system gives better clustering results than cluster forests for eight standard clustering benchmarks from UC Irvine Machine Learning Repository.

Due to the ever increasing amount and severity of attacks aimed at compromising smartphones in general, and Android devices in particular, much effort have been devoted in recent years to deal with such incidents. However, scant attention has been devoted to study the interplay between visualization techniques and Android malware detection. As an initial proposal, neural projection architectures are applied in present work to analyze malware apps data and characterize malware families. By the advanced and intuitive visualization, the proposed solution provides with an overview of the structure of the families dataset and ease the analysis of their internal organization. Dimensionality reduction based on unsupervised neural networks is performed on family information from the Android Malware Genome (Malgenome) dataset.

In data-communication networks, network reliability is of great concern to both network operators and customers. To provide network reliability it is fundamentally important to know the ongoing tasks in a network. A particular task may depend on multiple network services, spanning many network devices. Unfortunately, dependency details are often not documented and are difficult to discover by relying on human expert knowledge. In monitored networks huge amounts of data are available and by applying data mining techniques, we are able to extract information of ongoing network activities. Hence, we aim to automatically learn network dependencies by analyzing network traffic and derive ongoing tasks in data-communication networks. To automatically learn network dependencies, we propose a methodology based on the normalized form of cross correlation, which is a well-established methodology for detecting similar signals in feature matching applications.

A new chaotic, permutation-substitution architecture based, image encryption algorithm has been introduced with a novel inter-intra bit-level permutation based confusion strategy. The proposed image cryptosystem uses Sudoku grids to ensure a high performance diffusion process and key space. This paper presents a fully comprehensive set of security analyses on the newly proposed image cryptosystem, including statistical and differential analysis, local and global information entropy calculation and robustness profile to different types of attacks. Based on the good statistical results, rounded by the theoretical arguments, the conducted study demonstrates that the proposed images’ encryption algorithm has a desirable level of security.

Due to the ever increasing amount and severity of attacks aimed at compromising smartphones in general, and Android devices in particular, much effort have been devoted in recent years to deal with such incidents. However, accurate detection of bad-intentioned Android apps still is an open challenge. As a follow-up step in an ongoing research, preset paper explores the selection of features for the characterization of Android-malware families. The idea is to select those features that are most relevant for characterizing malware families. In order to do that, an evolutionary algorithm is proposed to perform feature selection on the Drebin dataset, attaining interesting results on the most informative features for the characterization of representative families of existing Android malware.

Developed initially for tasks related to computer graphics, GPUs are increasingly being used for general purpose processing, including scientific and engineering applications. In this contribution, we have analysed the performance of three graphics cards that belong to the parallel computing CUDA platform with two C++ and Java multi-threading implementations, using as an example of computation a brute-force attack on KeeLoq, one of the best known remote keyless entry applications. As it was expected, these implementations are not able to break algorithms with 64-bit keys, but the results allow us to provide valuable information regarding the compared capabilities of the tested platforms.

In 2015, Kumaraswamy et al. have proposed an improvement of the key authentication scheme based on discrete logarithms. That kind of schemes has been widely studied for many years, producing many modifications and improvements designed to overcome the weaknesses detected; most of them leading to key substitution attacks and, in some cases, allowing to recover the user’s private key. The improvement proposed by Kumaraswamy et al. is based on the Chinese remainder theorem in combination with the discrete logarithm. In this paper, several mathematical inconsistencies are revealed in the definition. Once fixed, a key substitution attack is performed.

The main goal of this work is to propose a novel mathematical model to simulate malware spreading in wireless networks considering carrier devices (those devices that malware has reached but it is not able to carry out its malicious purposes for some reasons: incompatibility of the host’s operative system with the operative system targeted by the malware, etc.) Specifically, it is a SCIRS model (Susceptible-Carrier-Infectious-Recovered-Susceptible) where reinfection and vaccination are considered. The dynamic of this model is studied determining the stability of the steady states and the basic reproductive number. The most important control strategies are determined taking into account the explicit expression of the basic reproductive number.

Mathematical models to study to simulate the spread of malware are widely studied today. Malware spreading in Wireless Sensor Networks (WSNs) has special relevance as these networks consist on hundreds or even thousands of autonomous devices (sensors) able to monitor and to communicate with one another. Malware attacks on WSNs have become a critical challenge because sensors generally have weak defense capabilities, that is why the malware propagation in WSNs is relevant for security community. In this paper, some of the most important and recent global mathematical models to describe malware spreading in such networks are presented.

Elliptic Curve Cryptography (ECC) is a branch of public-key cryptography based on the arithmetic of elliptic curves. In the short life of ECC, most standards have proposed curves defined over prime finite fields satisfying the curve equation in the short Weierstrass form. However, some researchers have started to propose as a more secure alternative the use of Edwards and Montgomery elliptic curves, which could have an impact in current ECC deployments. This contribution evaluates the performance of the three types of elliptic curves using some of the examples provided by the initiative SafeCurves and a Java implementation developed by the authors, which allows us to offer some conclusions about this topic.

This paper describes the utilization of epidemiological models, usually employed for malware propagation, to study the effects of random jamming attacks, which can affect the physical and MAC/link layers of all nodes in a wireless sensor network, regardless of the complexity and computing power of the devices. The random jamming term considers both the more classical approach of interfering signals, focusing on the physical level of the systems, and the cybersecurity approach that includes the attacks generated in upper layers, mainly in the MAC/link layer, producing the same effect on the communication channel. We propose, as a preliminary modelling task, the epidemiological mathematical model Susceptible–Exposed–Infected–Susceptible (SEIS), and analyze the basic reproductive number, the infection rate, the average incubation time and the average infection time.

The growing development of virtual learning platforms is boosting a new type of Big Data and of Big Data Stream, those ones that can be labeled as e-learning Big Data. These data, coming from different sources of Virtual Learning Environments, such as communications between students and instructors as well as pupils tests, require accurate analysis and mining techniques in order to retrieve from them fruitful insights. This paper analyzes the main features of current e-learning systems, pointing out their sources of data and the huge amount of information that may be retrieved from them. Moreover, we assess the concept of educational Big Data, suggesting a logical and functional layered model that can turn to be very useful in real life.

This work describes one application of the PBL (Project Based Learning) methodology at the curriculum of the “Industrial Informatics” module of the B.Sc. Degree in Industrial Electronics and Automation, taught at the University College of Engineering of Vitoria-Gasteiz. The choice of this methodology is based on the project orientation expected for the future engineers. Authors intended to reproduce, at reduced scale, the problematic of working in multidisciplinary teams with strict completion times. This approach forced students to get involved in the learning process while carrying out the tasks. During the process, students detected the learning needs by themselves in order to accomplish the project, and had to learn how to apply, in a proactive and autonomous way, the necessary techniques during the project implementation. Instructors proposed implementing a controller for a SCARA robot, which is a typical configuration found in industrial environments. Students take the “Industrial Informatics” module (first semester) previously to the “Robotics” module (second semester) of the third year, so the authors proposed a simplified configuration for the SCARA robot with only two degrees of freedom (2DoF). The scale model of the robot was built with the LEGO Mindstorms NXT kit, which provides an interesting flexibility/price compromise. This approach forced students to apply concepts acquired in previous modules, as well as skills and techniques that will be of use in the future.

This project describes the coordination undertaken in the Bachelor’s degree in Geomatic and Surveying Engineering in the University of the Basque Country (UPV/EHU) after the integration in the European Higher Education Area, explaining how cooperative learning can positively contribute to the teaching-learning process.

It is well known that there are high dropout rates in engineering faculties. There are plenty of reasons that can explain these bad results, among which we can mention group integration problems, lack of motivation or difficulty to adapt to the university study methodology. In this work we propose a Four-day Welcome Program for first-year students to help these new students to overcome some difficulties they will have to face during their time at the Engineering School. Our objective is to improve the results of the first year students by motivating them from the first day.

Hill climbing and simulated annealing are two fundamental search techniques integrating most artificial intelligence and machine learning courses curricula. These techniques serve as introduction to stochastic and probabilistic based metaheuristics. Simulated annealing can be considered a hill-climbing variant with a probabilistic decision. While simulated annealing is conceptually a simple algorithm, in practice it can be difficult to parameterize. In order to promote a good simulated annealing algorithm perception by students, a simulation experiment is reported here. Key implementation issues are addressed, both for minimization and maximization problems. Simulation results are presented.

In this paper, we make a pedagogical proposal to study the basic elements of circulation in architecture undergraduate degrees, using the Minecraft game as a tool. The theoretical basis for the proposal are Vygotsky’s sociointeractional theory and Ausubel’s theory of meaningful learning. We find ourselves reflecting on Information and Communication Technologies (ICT), gamification and video games in education. We outline some basic elements about circulation in buildings, its types, functionality, and accessibility, and the creativity needed to solve circulation problems in architecture. We introduce the Minecraft game, its characteristics, elements and use it as educational tool. We conclude that video games, specifically Minecraft, are of high interest in education as they develop skills for problem solving, collaborative work, research motivation and proactivity.

In this paper, an experience to approach the competence about ethical aspects of the profession is presented. Following an existing methodology, several cases are presented to the students in order to determine if people involved have had a professional or ethical behaviour. Codes of professional ethics or conduct have been also discussed with the students. The experience has been successful since students have actively participated and valued the methodology positively. This solves the lack of prior training in these ethical aspects.

This paper deals with a fuzzy expert system for evaluating teachers in e-learning. The special fuzzy expert system for evaluating teacher using satisfaction questionnaire with evaluative linguistic expressions and separate knowledge bases of expert system is proposed. Proposed expert system was created using Linguistic Fuzzy Logic Controller. In the proposed fuzzy expert system the theory of Natural Fuzzy Logic is applied.

Along the Science Week in San Sebastian, a workshop about quadrotors was presented where basic concepts of these machines and physical laws in which they are based were explained. This activity was framed in a workshop for high school pupils and families.

This paper introduces an ongoing Erasmus+ project granted in the KA2 “Cooperation for innovation and the exchange of good practices” call. The project was applied by the Gazi University (Turkey) and its duration is three years. It engages five universities and involves B.Sc., M.Sc. and Ph.D. students in the scope of renewable energies from different points of view, focusing on topics following the main interests of the partners.

In undergraduate classes it is important to understand theoretical aspects in relation with practical laboratory applications. Rasp-berry Pi is a single board computer that is inexpensive and can be used to showcase the intended behavior. Huffman encoding is used in lossless data compression and is a good example for educational applications that target any operating system and can be easily implemented in Java and C#. This paper presents the implementation of Huffman algorithm on multiple frameworks (Oracle Java, OpenJDK Java, .NET Framework, Mono Framework, .Net Core Framework) in order to test its performance for Linux and Windows IoT.

A networking hands-on laboratory involves connecting cables, configuring network equipment and testing the resulted network and its traffic. Virtualization, which emulates hardware with software, is used today to better use the hardware resources and offers flexibility when it comes to network connectivity design. This paper presents the challenges related to the use of type 1 and type 2 virtualization solutions in a university laboratory, focusing on the solutions Oracle VirtualBox and Microsoft Hyper-V.

Current smart devices (phones, tablets, etc.) have integrated light sensors to adjust the screen’s brightness to the ambient light. The light sensors have become more sensitive and are even able to read the RGB light components. In Android, this information can be accessed without special access rights for the application. An application can use the information from the light sensor to detect the ambient light variations and relay this information to a server where it can be used to determine the video information being displayed. This paper details the data flow and tests the implementation for a single video flow on multiple light sensors.

Knowing the advantages of public cloud services is important for undergraduate students that will work in the IT domain. The study of cloud technology should cover not only the theoretical aspects but should also give hands on access and experience with the management interface. Two important players on this market offer access to services as a free trial that allows testing and light cloud platform usage: Microsoft Azure and Google Cloud. This paper compares these free services and their usability in an undergraduate laboratory at the Computer Science specialization and proposes a laboratory structure that should cover this process of investigation.