Search results for: Large scale networks

Authors: Farhad Asadi, S. Hossein Sadati

Abstract:

This paper presents a prediction performance of feedforward Multilayer Perceptron (MLP) and Echo State Networks (ESN) trained with extended Kalman filter. Feedforward neural networks and ESN are powerful neural networks which can track and predict nonlinear signals. However, their tracking performance depends on the specific signals or data sets, having the risk of instability accompanied by large error. In this study we explore this process by applying different network size and leaking rate for prediction of nonlinear or chaotic signals in MLP neural networks. Major problems of ESN training such as the problem of initialization of the network and improvement in the prediction performance are tackled. The influence of coefficient of activation function in the hidden layer and other key parameters are investigated by simulation results. Extended Kalman filter is employed in order to improve the sequential and regulation learning rate of the feedforward neural networks. This training approach has vital features in the training of the network when signals have chaotic or non-stationary sequential pattern. Minimization of the variance in each step of the computation and hence smoothing of tracking were obtained by examining the results, indicating satisfactory tracking characteristics for certain conditions. In addition, simulation results confirmed satisfactory performance of both of the two neural networks with modified parameterization in tracking of the nonlinear signals.

Keywords: Feedforward neural networks, nonlinear signal prediction, echo state neural networks approach, leaking rates, capacity of neural networks.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 758

Authors: Gábor Szűcs, Gyula Sallai

Abstract:

The goal of this paper is to find Wardrop equilibrium in transport networks at case of uncertainty situations, where the uncertainty comes from lack of information. We use simulation tool to find the equilibrium, which gives only approximate solution, but this is sufficient for large networks as well. In order to take the uncertainty into account we have developed an interval-based procedure for finding the paths with minimal cost using the Dempster-Shafer theory. Furthermore we have investigated the users- behaviors using game theory approach, because their path choices influence the costs of the other users- paths.

Keywords: Dempster-Shafer theory, S-O and U-Otransportation network, uncertainty of information, Wardropequilibrium.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1530

Authors: Ciprian Dobre

Abstract:

The scale, complexity and worldwide geographical spread of the LHC computing and data analysis problems are unprecedented in scientific research. The complexity of processing and accessing this data is increased substantially by the size and global span of the major experiments, combined with the limited wide area network bandwidth available. We present the latest generation of the MONARC (MOdels of Networked Analysis at Regional Centers) simulation framework, as a design and modeling tool for large scale distributed systems applied to HEP experiments. We present simulation experiments designed to evaluate the capabilities of the current real-world distributed infrastructure to support existing physics analysis processes and the means by which the experiments bands together to meet the technical challenges posed by the storage, access and computing requirements of LHC data analysis within the CMS experiment.

Keywords: Modeling and simulation, evaluation, large scale distributed systems, LHC experiments, CMS.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1811

Authors: Wei Zhang

Abstract:

With the rapid development of deep learning, neural network and deep learning algorithms play a significant role in various practical applications. Due to the high accuracy and good performance, Convolutional Neural Networks (CNNs) especially have become a research hot spot in the past few years. However, the size of the networks becomes increasingly large scale due to the demands of the practical applications, which poses a significant challenge to construct a high-performance implementation of deep learning neural networks. Meanwhile, many of these application scenarios also have strict requirements on the performance and low-power consumption of hardware devices. Therefore, it is particularly critical to choose a moderate computing platform for hardware acceleration of CNNs. This article aimed to survey the recent advance in Field Programmable Gate Array (FPGA)-based acceleration of CNNs. Various designs and implementations of the accelerator based on FPGA under different devices and network models are overviewed, and the versions of Graphic Processing Units (GPUs), Application Specific Integrated Circuits (ASICs) and Digital Signal Processors (DSPs) are compared to present our own critical analysis and comments. Finally, we give a discussion on different perspectives of these acceleration and optimization methods on FPGA platforms to further explore the opportunities and challenges for future research. More helpfully, we give a prospect for future development of the FPGA-based accelerator.

Keywords: Deep learning, field programmable gate array, FPGA, hardware acceleration, convolutional neural networks, CNN.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 895

Authors: Mohammadreza Amini, Omid Moradtalab, Ebadollah Zohrevandi

Abstract:

In real-time networks a large number of application programs are relying on video data and heterogeneous data transmission techniques. The aim of this research is presenting a method for end-to-end vouch quality service in surface applicationlayer for sending video data in comparison form in wireless heterogeneous networks. This method tries to improve the video sending over the wireless heterogeneous networks with used techniques in surface layer, link and application. The offered method is showing a considerable improvement in quality observing by user. In addition to this, other specifications such as shortage of data load that had require to resending and limited the relation period length to require time for second data sending, help to be used the offered method in the wireless devices that have a limited energy. The presented method and the achieved improvement is simulated and presented in the NS-2 software.

Keywords: Heterogeneous wireless networks, adaptation mechanism, multi-level, Handoff, stop mechanism, graceful degrades, application layer.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1669

Authors: C. Benjbara, A. Habbani

Abstract:

Wireless networks are getting more and more used in every new technology or feature, especially those without infrastructure (Ad hoc mode) which provide a low cost alternative to the infrastructure mode wireless networks and a great flexibility for application domains such as environmental monitoring, smart cities, precision agriculture, and so on. These application domains present a common characteristic which is the need of coexistence and intercommunication between modules belonging to different types of ad hoc networks like wireless sensor networks, mesh networks, mobile ad hoc networks, vehicular ad hoc networks, etc. This vision to bring to life such heterogeneous networks will make humanity duties easier but its development path is full of challenges. One of these challenges is the communication complexity between its components due to the lack of common or compatible protocols standard. This article proposes a new patented routing protocol based on the OLSR standard in order to resolve the heterogeneous ad hoc networks communication issue. This new protocol is applied on a specific network architecture composed of MANET, VANET, and FANET.

Keywords: Ad hoc, heterogeneous, ID-Node, OLSR.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 749

Authors: Huang Shengqin, Xiao Hong

Abstract:

LES with mixed subgrid-scale model has been used to simulate aerodynamic performance of hypersonic configuration. The simulation was conducted to replicate conditions and geometry of a model which has been previously tested. LES Model has been successful in predict pressure coefficient with the max error 1.5% besides afterbody. But in the high Mach number condition, it is poor in predict ability and product 12.5% error. The calculation error are mainly conducted by the distribution swirling. The fact of poor ability in the high Mach number and afterbody region indicated that the mixed subgrid-scale model should be improved in large eddied especially in hypersonic separate region. In the condition of attach and sideslip flight, the calculation results have waves. LES are successful in the prediction the pressure wave in hypersonic flow.

Keywords: Hypersonic, LES, mixed Subgrid-scale model, experiment.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1572

Authors: S. Babaie-Kafaki, R. Ghanbari, S.H. Nasseri, E. Ardil

Abstract:

Bus networks design is an important problem in public transportation. The main step to this design, is determining the number of required terminals and their locations. This is an especial type of facility location problem, a large scale combinatorial optimization problem that requires a long time to be solved. The genetic algorithm (GA) is a search and optimization technique which works based on evolutionary principle of natural chromosomes. Specifically, the evolution of chromosomes due to the action of crossover, mutation and natural selection of chromosomes based on Darwin's survival-of-the-fittest principle, are all artificially simulated to constitute a robust search and optimization procedure. In this paper, we first state the problem as a mixed integer programming (MIP) problem. Then we design a new crossover and mutation for bus terminal location problem (BTLP). We tested the different parameters of genetic algorithm (for a sample problem) and obtained the optimal parameters for solving BTLP with numerical try and error.

Keywords: Bus networks, Genetic algorithm (GA), Locationproblem, Mixed integer programming (MIP).

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2305

Authors: Radouane Iqdour, Abdelouhab Zeroual

Abstract:

The Multi-Layered Perceptron (MLP) Neural networks have been very successful in a number of signal processing applications. In this work we have studied the possibilities and the met difficulties in the application of the MLP neural networks for the prediction of daily solar radiation data. We have used the Polack-Ribière algorithm for training the neural networks. A comparison, in term of the statistical indicators, with a linear model most used in literature, is also performed, and the obtained results show that the neural networks are more efficient and gave the best results.

Keywords: Daily solar radiation, Prediction, MLP neural networks, linear model

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1328

Authors: Dimitrios Koukopoulos

Abstract:

In this work, we study the impact of dynamically changing link slowdowns on the stability properties of packetswitched networks under the Adversarial Queueing Theory framework. Especially, we consider the Adversarial, Quasi-Static Slowdown Queueing Theory model, where each link slowdown may take on values in the two-valued set of integers {1, D} with D > 1 which remain fixed for a long time, under a (w, ¤ü)-adversary. In this framework, we present an innovative systematic construction for the estimation of adversarial injection rate lower bounds, which, if exceeded, cause instability in networks that use the LIS (Longest-in- System) protocol for contention-resolution. In addition, we show that a network that uses the LIS protocol for contention-resolution may result in dropping its instability bound at injection rates ¤ü > 0 when the network size and the high slowdown D take large values. This is the best ever known instability lower bound for LIS networks.

Keywords: Network stability, quality of service, adversarial queueing theory, greedy scheduling protocols.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1229

Authors: A. S. Tolba, A.H. El-Baz, A.A. El-Harby

Abstract:

The task of face recognition has been actively researched in recent years. This paper provides an up-to-date review of major human face recognition research. We first present an overview of face recognition and its applications. Then, a literature review of the most recent face recognition techniques is presented. Description and limitations of face databases which are used to test the performance of these face recognition algorithms are given. A brief summary of the face recognition vendor test (FRVT) 2002, a large scale evaluation of automatic face recognition technology, and its conclusions are also given. Finally, we give a summary of the research results.

Keywords: Combined classifiers, face recognition, graph matching, neural networks.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 7723

Authors: Dimitrios Koukopoulos

Abstract:

In this work, we study the impact of dynamically changing link slowdowns on the stability properties of packetswitched networks under the Adversarial Queueing Theory framework. Especially, we consider the Adversarial, Quasi-Static Slowdown Queueing Theory model, where each link slowdown may take on values in the two-valued set of integers {1, D} with D > 1 which remain fixed for a long time, under a (w, p)-adversary. In this framework, we present an innovative systematic construction for the estimation of adversarial injection rate lower bounds, which, if exceeded, cause instability in networks that use the LIS (Longest-in- System) protocol for contention-resolution. In addition, we show that a network that uses the LIS protocol for contention-resolution may result in dropping its instability bound at injection rates p > 0 when the network size and the high slowdown D take large values. This is the best ever known instability lower bound for LIS networks.

Keywords: Parallel computing, network stability, adversarial queuing theory, greedy scheduling protocols.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1415

Authors: Eman I. Raslan, Haitham S. Hamza, Reda A. El-Khoribi

Abstract:

Fiber-Wireless (FiWi) networks are a promising candidate for future broadband access networks. These networks combine the optical network as the back end where different passive optical network (PON) technologies are realized and the wireless network as the front end where different wireless technologies are adopted, e.g. LTE, WiMAX, Wi-Fi, and Wireless Mesh Networks (WMNs). The convergence of both optical and wireless technologies requires designing architectures with robust efficient and effective bandwidth allocation schemes. Different bandwidth allocation algorithms have been proposed in FiWi networks aiming to enhance the different segments of FiWi networks including wireless and optical subnetworks. In this survey, we focus on the differentiating between the different bandwidth allocation algorithms according to their enhancement segment of FiWi networks. We classify these techniques into wireless, optical and Hybrid bandwidth allocation techniques.

Keywords: Fiber-Wireless (FiWi), dynamic bandwidth allocation (DBA), passive optical networks (PON), media access control (MAC).

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2124

Authors: S. R. Shamili, J. Karthikeyan

Abstract:

Fast development of the total populace and far and wide urbanization has surprisingly expanded the advancement of the construction industry. As a result of these activities, old structures are being demolished to make new buildings. Due to these large-scale demolitions, a huge amount of debris is generated all over the world, which results in a landfill. The use of construction and demolition waste as landfill causes groundwater contamination, which is hazardous. Using construction and demolition waste as aggregate can reduce the use of natural aggregates and the problem of mining. The objective of this study is to provide a detailed overview on how the construction and demolition waste material has been used as aggregate in structural concrete. In this study, the preparation, classification, and composition of construction and demolition wastes are also discussed.

Keywords: Aggregate, construction and demolition waste, landfill, large scale demolition.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 643

Authors: Kati Kõrbe Kaare, Kristjan Kuhi, Ott Koppel

Abstract:

Transportation authorities need to provide the services and facilities that are critical to every country-s well-being and development. Management of the road network is becoming increasingly challenging as demands increase and resources are limited. Public sector institutions are integrating performance information into budgeting, managing and reporting via implementing performance measurement systems. In the face of growing challenges, performance measurement of road networks is attracting growing interest in many countries. The large scale of public investments makes the maintenance and development of road networks an area where such systems are an important assessment tool. Transportation agencies have been using performance measurement and modeling as part of pavement and bridge management systems. Recently the focus has been on extending the process to applications in road construction and maintenance systems, operations and safety programs, and administrative structures and procedures. To eliminate failure and dysfunctional consequences the importance of obtaining objective data and implementing evaluation instrument where necessary is presented in this paper

Keywords: Key performance indicators, performance measurement system, evaluation, system architecture.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2011

Authors: Hazem M. El-Bakry, Qiangfu Zhao

Abstract:

In recent years, fast neural networks for object/face detection have been introduced based on cross correlation in the frequency domain between the input matrix and the hidden weights of neural networks. In our previous papers [3,4], fast neural networks for certain code detection was introduced. It was proved in [10] that for fast neural networks to give the same correct results as conventional neural networks, both the weights of neural networks and the input matrix must be symmetric. This condition made those fast neural networks slower than conventional neural networks. Another symmetric form for the input matrix was introduced in [1-9] to speed up the operation of these fast neural networks. Here, corrections for the cross correlation equations (given in [13,15,16]) to compensate for the symmetry condition are presented. After these corrections, it is proved mathematically that the number of computation steps required for fast neural networks is less than that needed by classical neural networks. Furthermore, there is no need for converting the input data into symmetric form. Moreover, such new idea is applied to increase the speed of neural networks in case of processing complex values. Simulation results after these corrections using MATLAB confirm the theoretical computations.

Keywords: Fast Code/Data Detection, Neural Networks, Cross Correlation, real/complex values.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1626

Authors: S. Khanpour, A. Movaghar

Abstract:

The one of best robust search technique on large scale search area is heuristic and meta heuristic approaches. Especially in issue that the exploitation of combinatorial status in the large scale search area prevents the solution of the problem via classical calculating methods, so such problems is NP-complete. in this research, the problem of winner determination in combinatorial auctions have been formulated and by assessing older heuristic functions, we solve the problem by using of genetic algorithm and would show that this new method would result in better performance in comparison to other heuristic function such as simulated annealing greedy approach.

Keywords: Bids, genetic algorithm, heuristic, metaheuristic, simulated annealing greedy.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1787

Authors: Minsoo Lee, Julee Choi, Sookyung Song

Abstract:

The increasing interest on processing data created by sensor networks has evolved into approaches to implement sensor networks as databases. The aggregation operator, which calculates a value from a large group of data such as computing averages or sums, etc. is an essential function that needs to be provided when implementing such sensor network databases. This work proposes to add the DURING clause into TinySQL to calculate values during a specific long period and suggests a way to implement the aggregation service in sensor networks by applying materialized view and incremental view maintenance techniques that is used in data warehouses. In sensor networks, data values are passed from child nodes to parent nodes and an aggregation value is computed at the root node. As such root nodes need to be memory efficient and low powered, it becomes a problem to recompute aggregate values from all past and current data. Therefore, applying incremental view maintenance techniques can reduce the memory consumption and support fast computation of aggregate values.

Keywords: Aggregation, Incremental View Maintenance, Materialized view, Sensor Network.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1540

Authors: Gurjinder Kaur, Yogesh Chaba, V. K. Jain

Abstract:

The aim of this paper is to explore the security issues that significantly affect the performance of Mobile Adhoc Networks (MANET)and limit the services provided to their intended users. The MANETs are more vulnerable to Distributed Denial of Service attacks (DDoS) because of their properties like shared medium, dynamic topologies etc. A DDoS attack is a coordinated attempt made by malicious users to flood the victim network with the large amount of data such that the resources of the victim network are exhausted resulting in the deterioration of the network performance. This paper highlights the effects of different types of DDoS attacks in MANETs and categorizes them according to their behavior.

Keywords: Distributed Denial, Mobile Adhoc Networks

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2435

Authors: Vipin Pal, Girdhari Singh, R P Yadav

Abstract:

As wireless sensor networks are energy constraint networks so energy efficiency of sensor nodes is the main design issue. Clustering of nodes is an energy efficient approach. It prolongs the lifetime of wireless sensor networks by avoiding long distance communication. Clustering algorithms operate in rounds. Performance of clustering algorithm depends upon the round time. A large round time consumes more energy of cluster heads while a small round time causes frequent re-clustering. So existing clustering algorithms apply a trade off to round time and calculate it from the initial parameters of networks. But it is not appropriate to use initial parameters based round time value throughout the network lifetime because wireless sensor networks are dynamic in nature (nodes can be added to the network or some nodes go out of energy). In this paper a variable round time approach is proposed that calculates round time depending upon the number of active nodes remaining in the field. The proposed approach makes the clustering algorithm adaptive to network dynamics. For simulation the approach is implemented with LEACH in NS-2 and the results show that there is 6% increase in network lifetime, 7% increase in 50% node death time and 5% improvement over the data units gathered at the base station.

Keywords: Wireless Sensor Network, Clustering, Energy Efficiency, Round Time.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1786

Authors: Yiqin Lin, Liang Bao

Abstract:

This paper is devoted to the numerical solution of large-scale linear ill-posed systems. A multilevel regularization method is proposed. This method is based on a synthesis of the Arnoldi-Tikhonov regularization technique and the multilevel technique. We show that if the Arnoldi-Tikhonov method is a regularization method, then the multilevel method is also a regularization one. Numerical experiments presented in this paper illustrate the effectiveness of the proposed method.

Keywords: Discrete ill-posed problem, Tikhonov regularization, discrepancy principle, Arnoldi process, multilevel method.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 761

Authors: Chompunut Jantarasorn, Chutima Prommak

Abstract:

This paper presents an intensive measurement studying of the network performance analysis when IEEE 802.11g Wireless Local Area Networks (WLAN) coexisting with IEEE 802.15.4 Wireless Personal Area Network (WPAN). The measurement results show that the coexistence between both networks could increase the Frame Error Rate (FER) of the IEEE 802.15.4 networks up to 60% and it could decrease the throughputs of the IEEE 802.11g networks up to 55%.

Keywords: Wireless performance analysis, Coexistence analysis, IEEE 802.11g, IEEE 802.15.4.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1943

Authors: Samia Salah, M’hamed Hadj Sadok, Abderrezak Guessoum

Abstract:

In this paper, we propose a new modular approach called neuroglial consisting of two neural networks slow and fast which emulates a biological reality recently discovered. The implementation is based on complex multi-time scale systems; validation is performed on the model of the asynchronous machine. We applied the geometric approach based on the Gerschgorin circles for the decoupling of fast and slow variables, and the method of singular perturbations for the development of reductions models.

This new architecture allows for smaller networks with less complexity and better performance in terms of mean square error and convergence than the single network model.

Keywords: Gerschgorin’s Circles, Neuroglial Network, Multi time scales systems, Singular perturbation method.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1605

Authors: Hsun Su, Yuan-Kang Shih, Shin-Shin Kao

Abstract:

Many well-known interconnection networks, such as kary n-cubes, recursive circulant graphs, generalized recursive circulant graphs, circulant graphs and so on, are shown to belong to the family of cycle composition networks. Recently, various studies about mutually independent hamiltonian cycles, abbreviated as MIHC-s, on interconnection networks are published. In this paper, using an improved construction method, we obtain MIHC-s on cycle composition networks with a much weaker condition than the known result. In fact, we established the existence of MIHC-s in the cycle composition networks and the result is optimal in the sense that the number of MIHC-s we constructed is maximal.

Keywords: Hamiltonian cycle, k-ary n-cube, cycle composition networks, mutually independent.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1389

Authors: Zülküf Genç, Öznur Özkasap

Abstract:

Characteristics of ad hoc networks and even their existence depend on the nodes forming them. Thus, services and applications designed for ad hoc networks should adapt to this dynamic and distributed environment. In particular, multicast algorithms having reliability and scalability requirements should abstain from centralized approaches. We aspire to define a reliable and scalable multicast protocol for ad hoc networks. Our target is to utilize epidemic techniques for this purpose. In this paper, we present a brief survey of epidemic algorithms for reliable multicasting in ad hoc networks, and describe formulations and analytical results for simple epidemics. Then, P2P anti-entropy algorithm for content distribution and our prototype simulation model are described together with our initial results demonstrating the behavior of the algorithm.

Keywords: Ad hoc networks, epidemic, peer-to-peer, reliablemulticast.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1771

Authors: Yitao Lei, Xingxiang Zhai, Burra Venkata Durga Kumar

Abstract:

As the quantity and complexity of computing in large-scale software systems increase, distributed system computing becomes increasingly important. The distributed system realizes high-performance computing by collaboration between different computing resources. If there are no efficient resource scheduling resources, the abuse of distributed computing may cause resource waste and high costs. However, resource scheduling is usually an NP-hard problem, so we cannot find a general solution. However, some optimization algorithms exist like genetic algorithm, ant colony optimization, etc. The large scale of distributed systems makes this traditional optimization algorithm challenging to work with. Heuristic and machine learning algorithms are usually applied in this situation to ease the computing load. As a result, we do a review of traditional resource scheduling optimization algorithms and try to introduce a deep reinforcement learning method that utilizes the perceptual ability of neural networks and the decision-making ability of reinforcement learning. Using the machine learning method, we try to find important factors that influence the performance of distributed system computing and help the distributed system do an efficient computing resource scheduling. This paper surveys the application of deep reinforcement learning on distributed system computing resource scheduling. The research proposes a deep reinforcement learning method that uses a recurrent neural network to optimize the resource scheduling. The paper concludes the challenges and improvement directions for Deep Reinforcement Learning-based resource scheduling algorithms.

Keywords: Resource scheduling, deep reinforcement learning, distributed system, artificial intelligence.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 495

Authors: Nikhil, Bestamin Özkaya, Ari Visa, Chiu-Yue Lin, Jaakko A. Puhakka, Olli Yli-Harja

Abstract:

The performance of a sucrose-based H2 production in a completely stirred tank reactor (CSTR) was modeled by neural network back-propagation (BP) algorithm. The H2 production was monitored over a period of 450 days at 35±1 ºC. The proposed model predicts H2 production rates based on hydraulic retention time (HRT), recycle ratio, sucrose concentration and degradation, biomass concentrations, pH, alkalinity, oxidation-reduction potential (ORP), acids and alcohols concentrations. Artificial neural networks (ANNs) have an ability to capture non-linear information very efficiently. In this study, a predictive controller was proposed for management and operation of large scale H2-fermenting systems. The relevant control strategies can be activated by this method. BP based ANNs modeling results was very successful and an excellent match was obtained between the measured and the predicted rates. The efficient H2 production and system control can be provided by predictive control method combined with the robust BP based ANN modeling tool.

Keywords: Back-propagation, biohydrogen, bioprocessmodeling, neural networks.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1773

Authors: Dai Miyamoto

Abstract:

This paper examines the factors, which determine R&D outsourcing behaviour at Japanese firms, from the viewpoints of transaction cost and strategic management, since the latter half of the 1990s. This study uses empirical analysis, which involves the application of large-sample data. The principal findings of this paper are listed below. Firms that belong to a wider corporate group are more active in executing R&D outsourcing activities. Diversification strategies such as the expansion of product and sales markets have a positive effect on the R&D outsourcing behaviour of firms. Moreover, while quantitative R&D resources have positive influences on R&D outsourcing, qualitative indices have no effect. These facts suggest that R&D outsourcing behaviour of Japanese firms are consistent with the two perspectives of transaction cost and strategic management. Specifically, a conventional corporate group network plays an important role in R&D outsourcing behaviour. Firms that execute R&D outsourcing leverage 'old' networks to construct 'new' networks and use both networks properly.

Keywords: Corporate Group Networks, R&D Outsourcing, Strategic Management Perspective, Transaction Cost Perspective.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1682

Authors: S. Bouhouche, M. Lahreche, S. Ziani, J. Bast

Abstract:

Modern manufacturing facilities are large scale, highly complex, and operate with large number of variables under closed loop control. Early and accurate fault detection and diagnosis for these plants can minimise down time, increase the safety of plant operations, and reduce manufacturing costs. Fault detection and isolation is more complex particularly in the case of the faulty analog control systems. Analog control systems are not equipped with monitoring function where the process parameters are continually visualised. In this situation, It is very difficult to find the relationship between the fault importance and its consequences on the product failure. We consider in this paper an approach to fault detection and analysis of its effect on the production quality using an adaptive centring and scaling in the pickling process in cold rolling. The fault appeared on one of the power unit driving a rotary machine, this machine can not track a reference speed given by another machine. The length of metal loop is then in continuous oscillation, this affects the product quality. Using a computerised data acquisition system, the main machine parameters have been monitored. The fault has been detected and isolated on basis of analysis of monitored data. Normal and faulty situation have been obtained by an artificial neural network (ANN) model which is implemented to simulate the normal and faulty status of rotary machine. Correlation between the product quality defined by an index and the residual is used to quality classification.

Keywords: Modeling, fault detection and diagnosis, parameters estimation, neural networks, Fault Detection and Diagnosis (FDD), pickling process.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1577

Authors: Agung Budi Muljono, I Made Ginarsa, I Made Ari Nrartha

Abstract:

A large-scale power system (LSPS) consists of two or more sub-systems connected by inter-connecting transmission. Loading pattern on an LSPS always changes from time to time and varies depend on consumer need. The serious instability problem is appeared in an LSPS due to load fluctuation in all of the bus. Adaptive neuro-fuzzy inference system (ANFIS)-based power system stabilizer (PSS) is presented to cover the stability problem and to enhance the stability of an LSPS. The ANFIS control is presented because the ANFIS control is more effective than Mamdani fuzzy control in the computation aspect. Simulation results show that the presented PSS is able to maintain the stability by decreasing peak overshoot to the value of −2.56 × 10−5 pu for rotor speed deviation Δω2−3. The presented PSS also makes the settling time to achieve at 3.78 s on local mode oscillation. Furthermore, the presented PSS is able to improve the peak overshoot and settling time of Δω3−9 to the value of −0.868 × 10−5 pu and at the time of 3.50 s for inter-area oscillation.

Keywords: ANFIS, large-scale, power system, PSS, stability enhancement.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1194