Computer Science Journal (AGH University of Science and Technology, Krakow)
Not a member yet
476 research outputs found
Sort by
Influence of initial water saturation in earthen levees on results of numerical modelling of infiltration processes
Get PDFLevees in Poland are mostly earthen constructions. The stability of earthen levees depends largely on factors such as the construction material, meteorological conditions and natural elements. The influence of initial water saturation of pore space on levee stability is analysed in this paper. Analysis was performed using numerical modelling and water pore pressure results were compared against data obtained from sensors located in a levee. The numerical modelling shows the moderate influence of initial water level on distribution of water pore pressure during high water level
Analysis of Series of Measurements from Job-Centric Monitoring by Statistical Functions
Get PDFThe rising number of executed programs (jobs) enabled by thegrowing amount of available resources from Clouds, Grids,and HPC (for example) has resulted in an enormous number ofjobs. Nowadays, most of the executed jobs are mainlyunobserved, so unusual behavior, non-optimal resource usage,and silent faults are not systematically searched andanalyzed. Job-centric monitoring enables permanent jobobservation and, thus, enables the analysis of monitoringdata. In this paper, we show how statistic functions can beused to analyze job-centric monitoring data and how themethods compare to more-complex analysis methods.Additionally, we present the usefulness of job-centricmonitoring based on practical experiences
From Linear Classifier to Convolutional Neural Network for Hand Pose Recognition
Get PDFRecently gathered image datasets and new capabilities of high performance computing systems allowed developing new artificial neural network models and training algorithms. Using the new machine learning models, computer vision tasks can be accomplished based on the raw values of image pixels, instead of specific features. The principle of operation of deep artificial neural networks is more and more resembling of what we believe to be happening in the human visual cortex. In this paper we build up an understanding of convolutional neural networks through investigating supervised machine learning methods suchas K-Nearest Neighbors, linear classifiers and fully connected neural networks. We provide examples and accuracy results based on our implementation aimed for the problem of hand pose recognition
Notification Methods in Wireless Systems
Get PDFRecently, there has been an increasing need for secure, efficient, and simplenotification methods for wireless systems. Such systems are meant to provideusers with precise tools best suited for work or leisure environments, and a lotof effort has been put into creating a multitude of applications. At the sametime, however, not much research has been made into determining which of theavailable protocols are best suited for each individual task. A number of basicnotification methods are presented here, and tests have been performed for themost-promising ones. An attempt has been made to determine which of themethods have the best throughput, latency, security, and other characteristics.A comprehensive comparison is provided, which can be used to select the rightmethod for each individual project. Finally, conclusions are provided, and theresults from all of the tests conducted are discussed
Type-driven Development of Concurrent Communicating Systems
Get PDFModern software systems rely on communication, for example mobile applcations communicating with a central server, distributed systems coordinaing a telecommunications network, or concurrent systems handling events and processes in a desktop application. However, reasoning about concurrent prgrams is hard, since we must reason about each process and the order in which communication might happen between processes. In this paper, I describe a type-driven approach to implementing communicating concurrent programs, using the dependently typed programming language Idris. I show how the type system can be used to describe resource access protocols (such as controlling access to a file handle) and verify that programs correctly follow those prtools. Finally, I show how to use the type system to reason about the order of communication between concurrent processes, ensuring that each end of a communication channel follows a defined protocol
Smart levee in Poland. Full-scale monitoring experimental study of levees by different methods
Get PDFThis paper presents two types of control and measurement networks used in the levee built as part of the ISMOP project. The first control and measurement network based on pore pressure and temperature sensors. Additionally, it contains the fiber-optic technology. The second network include design experimental sensors, constructed for the development of solutions that can be used in existing flood embankments
Image Acquisition System based on Synchronized High Resolution Gigabit Ethernet Cameras
Get PDFOver the last few years, the huge rise in various computer vision applications canbe observed. They are widely used in such areas like video surveillance, medicaldiagnostics, biometrics recognition, the automotive or military industries. Mostof these solutions take advantage of high-resolution cameras in order to obtainhigh quality images. Surprisingly, little attention is paid in the literature tothe practical implementation of off-the-shelf image acquisition systems. Mostavailable solutions are composed of custom developed electronic devices whichuse specialized multi-core DSPs and / or FPGA technology. Therefore, in thispaper the novel realization of the scalable and comprehensive image acquisitionsystem based on synchronized high resolution Gigabit Ethernet camerasis presented. The proposed solution allows the connection of multiple camerastogether with any number of external illumination modules. Selected devicescan be synchronized with each other in user-defined configurations. Hence,designed solution can be easily integrated in both simple and complex applications.Authors describe in detail design and implementation processes of theproposed platform. The performance issues that can occur in such systems arepresented and discussed. Obtained results are encouraging and useful for thedevelopment of similar solutions
Predicting Performance in a PaaS Environment: a Case Study for a Web Application
Get PDFThis paper demonstrates how the combination of simulation and real-world experiments can be used to aid decisions concerning the performance of a distributed application. It presents a case study of performance analysis carried out for a commercial application implementing a web-based API server for mobile clients. The application was deployed on the Heroku cloud-based Platform as a Service (PaaS). The analysis described in this paper provided information required to choose the proper configuration of resources for the software. Simulation was used in the research to identify factors crucial to the performance of the application. This allowed for the preparation of basic experiments concentrating on these factors. Consequently, the basic parameters of resources crucial for the efficiency of the application could be benchmarked at insignificant cost and effort. This approach allows us to reliably aid decisions concerning resource configuration for an analyzed application. The simulation method used in this research is based on the formalism of Timed Colored Petri Nets, but the complexity of formal modeling is hidden from its users. Application developers are able to conveniently create a high-level model of their designs and perform simulations, while the reliability of the results is ensured by the formalism. The paper demonstrates the usefulness of the simulation method for analyzing real-world distributed systems
An Octopus-Inspired Intrusion Deterrence Model in Distributed Computing System
Get PDFThe study formulated and evaluated a model for effective management of maliciousnodes in mobile Ad-hoc network based on Ad-Hoc on- demand distancevector routing protocol. A collaborative injection model called CollaborativeInjection Deterrence Model (CIDM) was formulated using stochastic theory.The definition of the model was presented using graph theory. CIDM wassimulated using three different scenarios. The three scenarios were then comparedusing packets delivery ratio (PDR), routing load, throughput and delayas performance metrics. The simulation result showed that CIDM reduce considerablythe rate of packets dropped caused by malicious nodes in MANETnetwork. CIDM did not introduce additional load to the network and yet withproduce higher throughput. Lastly, the access delay with CIDM is minimalcompared with convectional OADV. The study developed a model to mete outa punitive measure to rogue nodes as a form of intrusion deterrence withoutdegrading the overall performance of the network. The well known CRAWDADdataset was used in the simulation
Reasoning Algorithm for a Creative Decision Support System Integrating Inference and Machine Learning
Get PDFIn this paper a reasoning algorithm for a creative decision support system is proposed. It allows to integrate inference and machine learning algorithms. Execution of learning algorithm is automatic because it is formalized as aplying a complex inference rule, which generates intrinsically new knowledge using the facts stored already in the knowledge base as training data. This new knowledge may be used in the same inference chain to derive a decision. Such a solution makes the reasoning process more creative and allows to continue resoning in cases when the knowledge base does not have appropriate knowledge explicit encoded. In the paper appropriate knowledge representation and infeence model are proposed. Experimental verification is performed on a decision support system in a casting domain