Skip to main content
main-content
Top

About this book

This book constitutes the proceedings of the 19th International Conference on Computer Information Systems and Industrial Management Applications, CISIM 2020, held in Bialystok, Poland, in October 2020.

The 40 full papers presented together with 5 abstracts of keynotes were carefully reviewed and selected from 62 submissions. The main topics covered by the chapters in this book are biometrics, security systems, multimedia, classification and clustering, industrial management. Besides these, the reader will find interesting papers on computer information systems as applied to wireless networks, computer graphics, and intelligent systems. The papers are organized in the following topical sections: biometrics and pattern recognition applications; computer information systems and security; industrial management and other applications; machine learning and high performance computing; modelling and optimization.

Table of Contents

Frontmatter

Biometrics and Pattern Recognition Applications

Frontmatter

Transfer Learning Approach in Classification of BCI Motor Imagery Signal

The paper presents application of a transfer learning-based, deep neural network classification model to the brain-computer interface EEG data. The model was initially trained on the publicly available dataset of motor imagery EEG data gathered from BCI experienced users. The final fitting was performed on the set of six participants for whom it was the first contact with a BCI system. The results show that initial training affects classification accuracy positively even in case of inexperienced participants. In the presented preliminary study five participants were examined. Data from each participant were analysed separately. Results show that the transfer learning approach allows to improve classification accuracy by even more than 10% points in comparison to the baseline deep neural network models, trained without transfer learning.

Filip Begiełło, Mikhail Tokovarov, Małgorzata Plechawska-Wójcik

Time Removed Repeated Trials to Test the Quality of a Human Gait Recognition System

The field of biometrics is currently an area that is both very interesting as well as rapidly growing. Among various types of behavioral biometrics, human gait recognition is worthy of particular attention. Unfortunately, one issue which is frequently overlooked in subject-related literature is the problem of the changing quality of a biometric system in relation to tests that are repeated after some time. The present article describes tests meant to assess the accuracy of a human gait recognition system based on Ground Reaction Forces in time removed repeated trials. Both the initial testing as well as the repeated trials were performed with the participation of the same 40 people (16 women and 24 men) which allowed the recording of nearly 1,600 stride sequences (approximately 800 in each trial). Depending on the adopted scenario correct recognition ranged from 90.4% to 100% of cases. These results indicate that the biometric system had greater problems with recognition the longer the period of time which passed since the first trials. The present article also analyzed the impact of footwear change in the second series of testing on recognition results.

Marcin Derlatka

Spiral-Based Model for Software Architecture in Bio-image Analysis: A Case Study in RSV Cell Infection

The advancement in biological and medical image acquisitions has allowed the development of numerous investigations in different fields supported by image analysis, from cell to physiological level. The complexity in the treatment of data, generated by image analysis, requires a structured methodology for software development. In this paper we proposed a framework to develop a software solution with a Service-Oriented Architecture (SOA) applied to the analysis of biological images. The framework is completed with a novel image analysis methodology that would help researchers to achieve better results in their image analysis projects. We evaluate our proposal in a scientific project related to cell image analysis.

Margarita Gamarra, Eduardo Zurek, Wilson Nieto, Miguel Jimeno, Deibys Sierra

Artificial Intelligence System for Drivers Fatigue Detection

Driver drowsiness is one of major causes of growing number of road accidents. To combat this problem car makers install their proprietary and expensive driver alert systems. In this paper an analogous system based on an opensource machine learning library is presented. The proper eye aspect ratio for closed eyes is discussed and estimated. The accuracy of closed eyes recognition is tested on a basis of several public available face libraries.

Waldemar Karwowski, Przemysław Reszke, Marian Rusek

Automatic Marking of Allophone Boundaries in Isolated English Spoken Words

The work presents a method that allows delimiting the borders of allophones in isolated English words. The described method is based on the DTW algorithm combining two signals, a reference signal and an analyzed one. As the reference signal, recordings from the MODALITY database were used, from which the words were extracted. This database was also used for tests, which were described. Test results show that the automatic determination of the allophone limits in English words is possible with good accuracy. Tests have been carried out to determine the error of particular allophones borders marking and to find out the cost of matching the given allophone to the reference one. Based on this cost, a coefficient has been introduced that allows for determining in percentage how much the automatically marked allophone is similar to the reference one. This coefficient can be used for an assessment of the correctness of the pronunciation of the allophone. The possibilities of further research and development of this method were also analyzed.

Janusz Rafałko, Andrzej Czyżewski

Computer Information Systems and Security

Frontmatter

Combined State Splitting and Merging for Implementation of Fast Finite State Machines in FPGA

A new method of the synthesis of finite state machines is proposed. In this method, the speed of FSM is taken into account already at the early stage of synthesis process. The method is based on sequential merging and splitting two internal states regarding to speed of FSM. This parameter may decrease with reduction of internal states, but splitting internal states leads to decrease of number of variables in logic functions which describe combinational part of FSM. This parameter has a great influence on a critical delay path. The results of experiments showing efficiency of proposed approach are also presented.

Adam Klimowicz

Securing Event Logs with Blockchain for IoT

The Internet of Things (IoT) is growing in popularity in recent years. With the increasing use, security threats are also becoming a bigger concern, especially considering new challenges, like limited computational power, low storage capacity and unprecedented number of independent, uncoordinated hardware device manufacturers. Unfortunately, modern attacks are more and more sophisticated and some attackers may even use anti-forensics techniques to hide any evidence of their malicious activity. As a result, digital forensics will be crucial in investigating crimes committed against IoT devices. One of the challenges is to create secure, lightweight and tamper-proof event log for the IoT system. Proposed solution relies on a blockchain to store event logs, guaranteeing the integrity of data. Event logs from IoT devices are being sent to multiple servers using multiple channels of communication to ensure logs availability and to move most of computational effort from the IoT device to the server. Logs are (optionally) encrypted to provide confidentiality of stored data. A set of security and performance tests were performed to prove effectiveness of proposed solution.

Mateusz Kłos, Imed El Fray

Securing Data of Biotechnological Laboratories Using Blockchain Technology

A few years ago blockchain technology was used in cryptocurrency. Nowadays, a variety of diverse areas are seeing the benefits of applying this technological approach to their needs. One way transactions without reverse mode is making blockchain a desirable platform for maintaining data. The authenticity, transparency and authorization of it make it ideal for healthcare or laboratory data systems. Research data should be authorized by a specific employee and never changed. The information collected should be never forged or falsified. Designing a blockchain system with access and permission rules is ideal in such a situation. In this article, we present the adaptation of blokchain technology to medical research laboratories and diagnostics. Afterwards, an RSA signature user can store information of any type and size. The new SHA-3 hash function is used to bind blocks together. This technological path makes laboratory workflow more efficient and fulfills restriction on medical laws.

Krzysztof Misztal, Tomasz Służalec, Aleksandra Kubica-Misztal

The Synthesis Method of High-Speed Finite State Machines in FPGA

The synthesis method of high-speed finite state machines (FSMs) in field programmable gate arrays (FPGAs) based on LUT (Look Up Table) by internal state splitting is offered. Estimations of the number of LUT levels are presented for an implementation of FSM transition functions in the case of sequential and parallel decomposition. Split algorithms of FSM internal states for the synthesis of high-speed FSMs are described. The method can be easily included in designing the flow of digital systems in FPGA. The experimental results showed a high efficiency of the offered method. FSM performance increased by 1.73 times. In conclusion, the experimental results were considered, and prospective directions for designing high-speed FSMs are specified.

Valery Salauyou, Damian Borecki, Tomasz Grzes

Industrial Management and other Applications

Frontmatter

A Framework of Business Intelligence System for Decision Making in Efficiency Management

The business decisions at different levels require processing different kinds of information. In this regard, the usage of suitable tools will contribute to making effective business decisions. The described framework of the business intelligence system aims to support such decisions in an effective way. The core of the proposed decision support system relies on several modules with a different database. One of them contains the input data of the particular problem, second include multi-criteria design analysis models, while the next contains optimization models to support decision-making. These optimization models are the focus of the current article. Two single and one multi-objective optimization models are formulated to express different situations and to support business decisions via reasonable solutions. Depending on the particular purpose, one of the models can be used to determine the best or compromise decision, which contributes to the effectiveness in business management. The applicability of the proposed models and respectively the core of the framework of business decision-making in efficiency management is illustrated in public street lights renovation. The obtained results show that all models are practically applicable in the determination of corresponding decisions in accordance with the selected goal. As the essences of the proposed framework are the optimization models this proves the effectiveness of optimization models in decision making to support efficient management.

Daniela Borissova, Petya Cvetkova, Ivan Garvanov, Magdalena Garvanova

Generalized Approach to Support Business Group Decision-Making by Using of Different Strategies

The recent advances in ICT and increased market competition make the problem of business decisions more significant and more complex. This is related to the performance evaluation of a variety of business decisions that have multi-level and multi-factor features. In this regard, the current article aims to propose a generalized flexible approach to support group decision making by using different strategies. The different decision-making strategies aim to provide the most preferable alternative; several good alternatives simultaneously, or ranking of all given alternatives. These different strategies are realized via corresponding optimization models that are capable to consider differences in the knowledge and expertise of the group experts. The contribution of the descried approach is focused on the aggregation stage of the known simple multi-attribute rating technique. The applicability of the proposed approach and formulated optimization models are demonstrated in the determination of preferable offer/s for printing a book considering several different evaluation parameters. The numerical results show that imposing requirements for different strategy realization it is quite helpful to get the group decision. Furthermore, the final group decision is modelled in such a way to reflect the particular background in expertise of each group’ member.

Daniela Borissova, Dilian Korsemov, Nina Keremedchieva

A Generic Materials and Operations Planning Approach for Inventory Turnover Optimization in the Chemical Industry

Chemical industries usually involve continuous and large-scale production processes that require demanding inventory control systems. This paper aims to show the results of the implementation of a mixed-integer programming model (MIP) based on the Generic Materials and Operations Planning Problem (GMOP) for optimizing the inventory turnover in a fertilizer company. Results showed significant improvements for Inventory Turnover Ratios and overall costs when compared with an empirical production planning method.

Jairo R. Coronado-Hernández, Alfonso R. Romero-Conrado, Olmedo Ochoa-González, Humberto Quintero-Arango, Ximena Vargas, Gustavo Gatica

Evolutionary Adaptation of (r, Q) Inventory Management Policy in Complex Distribution Systems

The paper addresses the inventory control problem in logistic networks with complex, mesh-type interconnection structure. Contrary to the majority of previously analyzed models, the considered topology does not assume any simplifications nor restrictions in the way the nodes are linked with each other. The system encompasses two types of actors – retailers and suppliers – connected via unidirectional links with non-negligible transshipment delay. The uncertain external demand may be imposed on any retailer and backordering is not allowed. The resource distribution is governed using the classical (r, Q) inventory management policy implemented in a distributed way. In this work, the continuous genetic algorithm is applied for automatic selection of reorder point r and shipment quantity Q. The optimization process aims to provide a trade-off between the economic costs and customer satisfaction. Numerous simulations are performed to evaluate the effectiveness of genetic algorithm performance in the considered class of problems.

Przemysław Ignaciuk, Łukasz Wieczorek

Design of a Decision Support System for Multiobjective Activity Planning and Programming Using Global Bacteria Optimization

The success of any project lies in a great manner on keeping costs in the estimated values, as well as meeting customer required due date. Therefore, there is a current need of developing an information system that facilitates the creation and managing of projects and their processes, including costing schemes, as well as monitoring an optimizing project’s makespan. In order to address this situation a user-friendly information system (IS) was developed. This IS includes an optimization module that reduces the project’s execution time, thus, minimizing costs and ultimately providing the manager with the right tools for the correct development of the project. Therefore, a better planning of activities in a reduced time is accomplished. In this way, the project manager is equipped with a decision support system (DSS) that allows a better decision making and, thanks to this performance optimization, a cost-effective solution can be delivered to the company. The optimization module is the main innovative component in this IS, considering that addresses the problem as a multiobjective one, considering at the same time makespan and cost. This module is based on global bacteria optimization (GBO). This becomes the most relevant improvement when compared to other ISs in the market.

Miguel Angel Jimenez-Barros, Diana Gineth Ramirez Rios, Carlos Julio Ardila Hernandez, Lauren Julieth Castro Bolaño, Dionicio Neira Rodado

Quality Improvement in Ammonium Nitrate Production Using Six Sigma Methodology

Six sigma has been used in different industries to reach operational excellence. However, in the chemical industry, the application of this methodology is limited. This research presents an implementation of the six sigma method for ammonium nitrate (AN) content optimization in condensate production for a fertilizer company in Colombia. The paper aims to determine the levels for input variables in the process, to meet desirable standards for condensate quality in terms of ammonium nitrate content. Based on the DMAIC steps implementation, it was possible to establish the main variables affecting the condensate quality and their optimal levels to reach an ammonium nitrate content below 15,000 ppm. These results demonstrate the impact that a six sigma project may have on operational effectiveness and quality improvement for meeting the customer requirements.

Olmedo Ochoa-González, Jairo R. Coronado-Hernández, Mayra A. Macías-Jiménez, Alfonso R. Romero-Conrado

Multicriteria Strategic Approach for the Selection of Concrete Suppliers in a Construction Company in Colombia

Within companies of the construction sector, the evaluation and selection of concrete suppliers, is considered a fundamental topic for the adequate construction of buildings, due to the impact it has, not only in the stability of the building but also in the productivity and the profitability. For an adequate selection, decision-makers must consider a series of criteria of different kind. However, as in different fields, due to the numerous criteria and alternatives that should be considered in the construction industry, the choice of an appropriate multicriteria decision-making approach has become a critical step in the selection of suppliers. Therefore, the objective of this research is to define the most adequate supplier of suitable concrete through the integration of powerful multicriteria decision-making methods. For this purpose, a fuzzy analytical hierarchy process (FAHP) is initially applied to define initial factor weights under uncertainty, followed, using the decision-making test and the evaluation laboratory (DEMATEL) to evaluate the interrelationships between elements of the hierarchy. Then, after combining FAHP and DEMATEL to calculate the final contributions of the factors including the interdependence, finally TOPSIS technique is used for the order of preference for the similarity with the ideal solution to evaluate and determine the best supplier.

Jorge E. Restrepo, Dionicio Neira Rodado, Amelec Viloria Silva

Machine Learning and High Performance Computing

Frontmatter

Representation Learning for Diagnostic Data

Representation learning algorithms have recently led to a significant progress in knowledge extraction from network structures. In this paper, a representation learning framework for the medical diagnosis domain is proposed. It is based on a heterogeneous network-based model of diagnostic data combined with an algorithm for learning latent node representation. Furthermore, a modification of metapath2vec algorithm is proposed for representation learning of heterogeneous networks. The proposed algorithm is compared with other representation learning approaches in two practical case studies: symptom/disease classification and disease prediction. A significant performance boost can be observed for these tasks, resulting from learning representations of domain data in a form of a heterogeneous network. It is also shown that in certain situations the modified algorithm improves the quality of learned embeddings compared to reference methods.

Karol Antczak

A Machine Learning Approach for Severe Maternal Morbidity Prediction at Rafael Calvo Clinic in Cartagena-Colombia

There is a huge problem in public health around the world called severe maternal morbidity (SMM). It occurs during pregnancy, delivery, or puerperium. This condition establishes risk for babies and women lives since it’s earlier detection isn’t easy [8]. In order to respond to such a situation, the current study suggests the use of logistic regression, and supports vector machine to construct a predicting model of risk level of maternal morbidity during pregnancy. Patients for the current study was the pregnant women who received prenatal care at Rafael Calvo Clinic in Cartagena, Colombia and final attention in the same clinic. This study presents the results of two machine learning algorithms, logistic regression and support vector machine. We validated the datasets from the first, second and third quarter of pregnancy with both techniques. The study shows that logistic regression achieves the best results with the prenatal control dataset from the first and second quarter and the support vector machine algorithm achieves the best prediction results with the data set from the third quarter. We generated two datasets using the information of medical records on pregnancy patients at Maternidad Rafael Calvo Clinic. The first dataset contains the six initial months of pregnancy data and the second dataset contains the last quarter of pregnancy data. We trained the first model with logistic regression and the datasets corresponding to the first semester of pregnancy. We obtained a classification of 97% sensibility, 51.8% positive predictive value and F1 score of 67.7%. The support vector machine model was implemented with the datasets obtained from the third quarter of pregnancy. We obtained a classifier with 100% of sensibility, 27.0% of precision.

Eugenia Arrieta Rodríguez, Fernando López-Martínez, Juan Carlos Martínez Santos

Collaborative Data Acquisition and Learning Support

With the constant development of neural networks, traditional algorithms relying on data structures lose their significance as more and more solutions are using AI rather than traditional algorithms. This in turn requires a lot of correctly annotated and informative data samples. In this paper, we propose a crowdsourcing based approach for data acquisition and tagging with support for Active Learning where the system acts as an oracle and repository of training samples. The paper presents the CenHive system implementing the proposed approach. Three different usage scenarios are presented that were used to verify the proposed approach.

Tomasz Boiński, Julian Szymański

Benchmarking Deep Neural Network Training Using Multi- and Many-Core Processors

In the paper we provide thorough benchmarking of deep neural network (DNN) training on modern multi- and many-core Intel processors in order to assess performance differences for various deep learning as well as parallel computing parameters. We present performance of DNN training for Alexnet, Googlenet, Googlenet_v2 as well as Resnet_50 for various engines used by the deep learning framework, for various batch sizes. Furthermore, we measured results for various numbers of threads with ranges depending on a given processor(s) as well as compact and scatter affinities. Based on results we formulate conclusions with respect to optimal parameters and relative performances which can serve as hints for researchers training similar networks using modern processors.

Klaudia Jabłońska, Paweł Czarnul

Binary Classification of Cognitive Workload Levels with Oculography Features

Assessment of cognitive workload level is important to understand human mental fatigue, especially in the case of performing intellectual tasks. The paper presents a case study on binary classification of cognitive workload levels. The dataset was received from two versions of the digit symbol substitution test (DSST), conducted on 26 healthy volunteers. A screen-based eye tracker was applied during an examination gathering oculographic data. DSST test results such as total number of matches and error ratio were also applied. Classification was performed with several different machine learning models. The best accuracy (97%) was achieved with linear SVM classifier. The final dataset for classification was based on nine features selected with the Fisher score feature selection method.

Monika Kaczorowska, Martyna Wawrzyk, Małgorzata Plechawska-Wójcik

Machine Learning Approach Applied to the Prevalence Analysis of ADHD Symptoms in Young Adults of Barranquilla, Colombia

Disorder Attention Deficit/Hyperactivity Disorder, or ADHD, is recognized as one of the pathologies of high prevalence in children and adolescents from the global environment population; this disorder generates visible symptoms usually diminish with the passage of time in adulthood, however they remain concealed by demonstrations damnifican personal stability and human development apt. This article shows the results of the research aimed at determining the prevalence of symptoms of attention deficit hyperactivity disorder in Young Adults University of Barranquilla and its Metropolitan Area. The sample of 1600 young adults between 18 and 25 years, which has been estimated at 95% confidence level and a margin of error of 2.44%. The information was acquired through the application of exploratory instruments symptoms of attention deficit hyperactivity disorder. With the application of the algorithm different machine learning algorithms such as: Bagging, MultiBoostAB, DecisionStump, LogitBoost, FT, J48Graft, a high performance in the Bagging algorithm could be identified with the following results in quality metrics: Accuracy 91.67%, Precision 94.12%, Recall 88.89% and F-measure 91.43%.

Alexandra Leon-Jacobus, Paola Patricia Ariza-Colpas, Ernesto Barcelo-Martínez, Marlon Alberto Piñeres-Melo, Roberto Cesar Morales-Ortega, David Alfredo Ovallos-Gazabon

Application of DenseNets for Classification of Breast Cancer Mammograms

In this study, we focus on the problem of a breast cancer diagnosis using mammography images by classifying them as belonging either to a negative or to a malignant mass class. We explore the potential of densely connected convolutional neural network (DenseNet) architectures by comparing its three different variants that were trained to classify the abnormalities in breast tissue. The models have been tested in a series of systematic experiments. With a limited dataset (2247 images per class), it was necessary to perform tests to verify whether the amount of data used in this work is sufficient to allow for the conclusion that the experimental results are not dependent on the subset of the data. The training was conducted using stratified 10-fold cross-validation to obtain statistically reliable metrics estimates. DenseNet-201 was found to be the best model achieving: 0.96 value for area under the curve (AUC), 0.92 for precision, 0.90 for recall, and 91% for accuracy.

Anita Rybiałek, Łukasz Jeleń

Augmentation of Segmented Motion Capture Data for Improving Generalization of Deep Neural Networks

This paper presents a method for augmenting the motion capture trajectories to improve generalization performance of recurrent long short-term memory (LSTM) neural networks. The presented algorithm is based on the interpolation of existing time series and can be applied only to segmented or easy-to-segment data due to the possibility of blending similar motion trajectories that are not significantly time-shifted. The paper shows the results of the classification efficiency with and without augmentation for two publicly available databases: Multimodal Kinect-IMU Dataset and National Chiao Tung University Multisensor Fitness Dataset. The former contains the data representing separate human computer interaction gestures, while the latter comprises the data of unsegmented series of body exercises. As a result of using the presented algorithm, the classification accuracy increased by approximately 11% points for the first dataset and 8% points for the second one.

Aleksander Sawicki, Sławomir K. Zieliński

Improving Classification of Basic Spatial Audio Scenes in Binaural Recordings of Music by Deep Learning Approach

The paper presents a deep learning algorithm for the automatic classification of basic spatial audio scenes in binaural music recordings. In the proposed method, the binaural audio recordings are initially converted to Mel-spectrograms, and subsequently classified using the convolutional neural network. The proposed method reached an accuracy of 87%, which constitutes a 10% improvement over the results reported in the literature. The method is capable of delivering moderate levels of classification accuracy even when single-channel spectrograms are directed to its input (e.g. solely from the left “ear”), highlighting the importance of monaural cues in spatial perception. The obtained results emphasize the significance of including multiple frequency bands in the convolution process. Visual inspection of the convolution filter activations reveals that the network performs a complex spectro-temporal sound decomposition, likely including a form of a foreground audio content separation from its background constituents.

Sławomir K. Zieliński

Modelling and Optimization

Frontmatter

AutoNet: Meta-model for Seamless Integration of Timed Automata and Colored Petri Nets

Time dependent modeling paradigm has always remained the focus of study for embedded system designers. The reason behind this rationale is that the safety and reliability of real time systems mainly depend on how precisely and accurately the time domain is modeled. Several time-driven models have been proposed and attained the level of maturity through series of developments. The most adopted formal methods are 1) Timed Automata which extends the finite states with finite number of real valued clocks, and 2) Colored Petri net which extends finite set of directed graphs with finite number of tokens coupled with color. In this paper, we proposed a Meta model (named as AutoNet) aimed at integration of timed automata and colored petri net. The main purpose of AutoNet is the transformation i.e. a single design with basic classes and state transition diagrams can be transformed to both timed automata and colored petri net. We performed case study to show proof of our concept prototype at traffic light signal modeling. A single iteration through the AutoNet produced both the timed automata and colored petri net of our test case, which is the validation of our design. AutoNet will serve as an automated ‘what you see is what you get’ (WYSIWYG) tool for embedded system engineers.

Muhammad Waqas Ahmad, Muhammad Waseem Anwar, Farooque Azam, Yawar Rasheed, Usman Ghani, Mukhtar Ahmad

A Multi-purpose Model Driven Platform for Contingency Planning and Shaping Response Measures

Effective emergency response requires situational awareness and preplanning for various contingencies inherent to the catastrophe; may it be a natural disaster or man-made crisis situation. Model Driven Software Engineering has contributed to the domain of contingency planning and response in a befitting manner by providing generic and scalable models, simulating emergency scenarios, in order to enhance the skills of responders. However, a thorough literature review identified that a comprehensive, intelligent and repository based model for planning of various contingencies inherent to the emergency situation is a mile stone to be achieved. Accepting the challenge, Interactive Contingency environment creation and Response Planning System (CRIPS) is proposed, which is a model driven platform/framework and facilitates a crisis manager to create a virtual emergency environment with its diverse ingredients and shaping an effective response actions to cater it. A multi perspective feedback mechanism and centrally administered picture board, which are the essential components of any response planning system, have also been incorporated. In addition, intelligent rater and repository concepts are introduced to rate the planning of crisis manager and save this whole contingency environment for analysis and self-learning of model, making it distinctive to the previous researches. The outcome of this research is a comprehensive meta-model, which can be further extended for model based development of an effective contingency and emergency response planning system. The validity of proposed meta-model is demonstrated through a real world case study of terrorists attack on Army Public School (APS) Peshawar/Pakistan. The results prove that proposed model is capable of modeling simple as well as complex scenarios and allows a crisis manager to effectively model a response in order to deter the catastrophic effects.

Mukhtar Ahmad, Farooque Azam, Yawar Rasheed, Muhammad Waseem Anwar, Muhammad Waqas Ahmad

Multi-criteria Differential Evolution for Optimization of Virtual Machine Resources in Smart City Cloud

In a smart city, artificial intelligence tools support citizens and urban services. From the user point of view, smart applications should bring computing to the edge of the cloud, closer to citizens with short latency. However, from the cloud designer point of view, the trade-off between cost, energy and time criteria requires the Pareto solutions. Therefore, the proposed multi-criteria differential evolution can optimize virtual machine resources in smart city clouds to find compromises between preferences of citizens and designers. In this class of distributed computer systems, smart mobile devices share computing workload with the set of virtual machines that can be migrated among the nodes of the cloud. Finally, some numerical results are studied for the laboratory cloud GUT-WUT.

Jerzy Balicki, Honorata Balicka, Piotr Dryja, Maciej Tyszka

Dynamic Ensemble Selection – Application to Classification of Cutting Tools

In order to improve pattern recognition performance of an individual classifier an ensemble of classifiers is used. One of the phases of creating the multiple classifier system is the selection of base classifiers which are used as the original set of classifiers. In this paper we propose the algorithm of the dynamic ensemble selection that uses median and quartile of correctly classified objects. The resulting values are used to define the decision schemes, which are used in the selection of the base classifiers process. The proposed algorithm has been verified on a real dataset regarding the classification of cutting tools. The obtained results clearly indicate that the proposed algorithm improves the classification measure. The improvement concerns the comparison with the ensemble of classifiers method without the selection.

Paulina Heda, Izabela Rojek, Robert Burduk

Stochastic Model of the Simple Cyber Kill Chain: Cyber Attack Process as a Regenerative Process

The proposed model extends the Markov model of the simple cyber kill chain already published in the literature by assumption of any continuous probability distribution of adversaries’ and defenders’ activity time. The description of the chain is based on the cyber kill chain concept initially introduced by Lockheed Martin’s researchers as the intrusion kill chain. The model includes the assumption of repeatability of cyber-attacks. On this basis, the stationary probabilities of staying of the attack process in individual phases were determined.

Romuald Hoffmann

Genetic Algorithm for Generation Multistage Tourist Route of Electrical Vehicle

The problem of selection points of interest which tourist wants to visit the most in the case when tourist travels by electric vehicle (EV) is examined in this paper. Furthermore the battery capacity of EVs is limited so charging stations are selected to the route, in order to a tourist could recharge the battery and move on to the next stage of the route. The genetic algorithm is proposed and tests for the different limitations of EVs batteries are conducted on realistic database. The experimental results show that the proposed genetic algorithm can successfully be used in this context.

Joanna Karbowska-Chilinska, Kacper Chociej

Event Ordering Using Graphical Notation for Event-B Models

System requirements are sometimes either too complex or undefined. Event-B is a formal modeling method and is being used increasingly to model various systems. Event-B models support atomicity decomposition and are quite useful for complex refinement structures. However, neither a Event-B model represents any explicit control flows among the events, nor does it support links between the new events during refinements. This work aims to model the Stop and Wait mechanism for an Automatic Repeat Request (ARQ) protocol to analyze the complexities due to communication errors during data re-transmissions. The limitation is the lack of control flows among the events during successive refinements. This has been graphically represented in this work and embedded with Event-B notations for the atomicity decomposition of the model. Finally, the successive refinements presented using an Event-B model, has been validated using the Rodin tool. This leads to a successful ARQ model.

Rahul Karmakar, Bidyut Biman Sarkar, Nabendu Chaki

Intraday Patterns in Trading Volume. Evidence from High Frequency Data on the Polish Stock Market

According to the literature, there are some possible shapes of intraday patterns in stock market characteristics such as trading volume, transaction costs, order flows, depths, spreads, price returns, stock market resiliency, etc. Empirical investigation and visualization of intraday patterns may be a useful tool for investment decision–making process and can help an analyst to state how particular characteristics vary over a session. In this paper, intraday patterns in trading volume based on high frequency data, are investigated. The data set is large, and it contains transaction data rounded to the nearest second for 10 companies traded in the Warsaw Stock Exchange (WSE). The whole sample covers the long period from January 2005 to December 2018. Extensive studies document various hour-of-the-day patterns in volume on the stock markets in the world. The findings of empirical experiments for real-data from the WSE are in general consistent with the literature and they confirm that intraday trading volume reveals U-similar or M-similar patterns in the case of all investigated equities, for all analyzed periods.

Joanna Olbryś, Adrian Oleszczak

An Efficient Metaheuristic for the Time-Dependent Team Orienteering Problem with Time Windows

The Time-Dependent Team Orienteering Problem with Time Windows (TDTOPTW) is a combinatorial optimization problem defined on graphs. The goal is to find most profitable set of paths in time-dependent graphs, where travel times (weights) between vertices varies with time. Its real life applications include tourist trip planning in transport networks. The paper presents an evolutionary algorithm with local search operators solving the problem. The algorithm was tested on public transport network of Athens and clearly outperformed other published methods achieving results close to optimal in short execution times.

Krzysztof Ostrowski

Measurement and Optimization Models of Risk Management System Usability

Systematization of issues related to the approach to defining and measuring the usability of a Risk Management System (RMS) is one of the key milestones on the way to understand it. A number of models and methods exists in this area. Author attempted to propose the model of the RMS usability and methodology of its application, as a results of in-depth studies carried out by himself in multiple organizations. The outcomes of research are aligned to the requirements that allow their use in any organization regardless of its field of operations, location, size and type of ownership. The model provides fundamentals that enables optimization of the RMS functional configuration in cases where it is possible to generate several acceptable its configuration’s variants for specific organization.

Tomasz Protasowicki

Development Methodology to Share Vehicles Optimizing the Variability of the Mileage

A simulation is a tool used to visualize the behaviors of a system, which will later help make decisions regarding how to handle the variables involved in the system, as well as the specific changes that have to be made. This study shows a case of vehicle allocation for different people within a company, evaluating methodologies, vehicle rotation to reduce the variance of the mileage and eliminating penalties with rental agencies for exceeding the permitted mileage. The paper shows a literature review of allocation models and similar studies, and later displays a detailed description of the problem, the variables that was used, the composition of the simulation and the optimization model that were generated, the results of the simulation, and finally, the findings of the research.

Luis E. Ramírez Polo, Alcides R. Santander-Mercado, Miguel A. Jimenez-Barros

Optimisation Model of Military Simulation System Maintenance

We present logistic support architecture of real simulation training system for military troops preparing to conduct battle operations at a level up to a whole battalion. Due to the need of cost reduction of system maintenance, the system components were designed to be easily serviced with an integrated supply chain support. A framework of multi-period multi-level multi-item production planning models with open and closed loop supply chains is introduced in the paper. Open loop supply chain (OLSC) and closed loop supply chain (CLSC) support alternatively production planning (PP) processes. In CLSC, a PP process is supported by the products recovery processes made in a closed loop chain. A computational complexity involved in preparing production plans for a long planning horizon is important problem when solving these models, so the model optimised for quick solution finding is presented. Most elements of the presented models are common for the military and civil industry.

Wojciech Stecz, Tadeusz Nowicki

Imbalanced Data: Rough Set Methods in Approximation of Minority Classes

The imbalanced data problem turned out to be one of the most important and challenging problems in artificial intelligence. We discuss an approach of minority class approximation based on rough set methods and three-way decision. This approach seems to be more general than the traditional one. However, it requires developing some new logical tools for reasoning based on rough sets and three-way decision, which is often expressed in natural language.

Jaroslaw Stepaniuk

Run-Time Schedule Adaptation Methods for Sensor Networks Coverage Problem

In this paper, we study run-time adaptation methods of a schedule of sensor activity generated for ideal temperature conditions. Such a schedule cannot be completed in low-temperature conditions due to a shorter lifetime of sensor batteries. We proposed several methods of selecting the next slot to be executed when the currently scheduled slot is unfeasible. Our experiments showed that in most cases, the best method is the one selecting the next slot based on computing the mean and standard deviation values of the battery load level for all the sensors active in a given slot.

Krzysztof Trojanowski, Artur Mikitiuk, Jakub A. Grzeszczak

Spectral Cluster Maps Versus Spectral Clustering

The paper investigates several notions of graph Laplacians and graph kernels from the perspective of understanding the graph clustering via the graph embedding into an Euclidean space. We propose hereby a unified view of spectral graph clustering and kernel clustering methods. The various embedding techniques are evaluated from the point of view of clustering stability (with respect to k-means that is the algorithm underpinning the spectral and kernel methods). It is shown that the choice of a fixed number of dimensions may result in clustering instability due to eigenvalue ties. Furthermore, it is shown that kernel methods are less sensitive to the number of used dimensions due to downgrading the impact of less discriminative dimensions.

Sławomir T. Wierzchoń, Mieczysław A. Kłopotek

Backmatter

Additional information

Premium Partner

    Image Credits