Scheduling Module Research Articles

KubFBS: A fine‐grained and balance‐aware scheduling system for deep learning tasks based on kubernetes

AbstractThe past decade witnessed a remarkable increase in deep learning (DL) workloads which require GPU resources to accelerate the training process. However, the existing coarse‐grained scheduling mechanisms are agnostic to information other than the number of GPUs or GPU memory, which results in performance degradation of DL tasks. Moreover, the common assumption held by the existing balance‐aware DL task scheduling strategies, a DL task consumes resources once it starts, fails to reduce resource contention, and further limits execution efficiency. To address these problems, this article proposes a fine‐grained and balance‐aware scheduling model (FBSM) which considers the resource consumption characteristic of the DL task. Based on FBSM, we propose customized GPU sniffer (GPU‐S) and balance‐aware scheduler (BAS) modules to construct a scheduling system called KubFBS. The experimental results demonstrate KubFBS accelerates the execution of DL tasks while improving the load balancing capability of the cluster.

An approach to dynamic malware analysis based on system and application code split

This paper discusses the development of tools for dynamic malware analysis. The main idea is to provide total control on a suspicious sample execution on the test computer. The approach we propose is to separate the application code from the system code by using memory pages access control. Thus, we are able to detect all system API calls and non-standard ways to transfer the control flow. Our tools (codename ToolChain) intentionally consist of a Control module, a Scheduling module, and a Cloaking module. In this paper, we focus mainly on the Control module. We monitor internal target process events by using invasive methods such as a system call hook or an executable file patch. This research describes the key creation stages of the prototype, with the basic functionality and technical ideas on handling several issues, such as analysis of multi-threaded applications, cloaking of the presence of analytical tools, and mitigation of the performance degradation of the operation system.

A Simulation Scheduling Module to Improve User Experience in the Simugan Beef-Cattle Farm Simulator

At the Faculty of Veterinary Sciences of the National University of Central Buenos Aires a client-server Beef-Cattle Farm simulator called Simugan has been developed. Simugan allows users to experiment over a virtual farm in a simple and low cost way compared with real farm conditions. Users can submit single simulation scenarios or multiple simulation scenarios packaged in an experimentation, where each scenario is a complete farm configuration. This is a key feature important in farm research, but with the drawback that some users might experiment long wait times for simulation results because of the amount of simulations the underlying hardware architecture has to process. Consequently, an heuristic scheduler module was added to Simugan producing a more equitative use of computer resources and an improvement of 41 % in users flow time, a popular metric to quantify how much time user simulations spend in the back end and hence a way of measuring deviations in users waiting times.

Blockchain-Based Intelligent Charging Station Management System Platform

A smart electric vehicle (EV) charging station energy management system (CSMS) based on blockchain technology, which aims to protect privacy of EV users, ensure fairness of power transactions, and meet charging demands for large numbers of EVs, is proposed in this study. EV charging pile is designed as a local blockchain distributed ledger node, which operates synchronously with blockchain system and blockchain distributed ledger in cloud server. This paper integrates CSMS through smart contracts, providing EV users that ability to conduct power transactions and perform optimal charging and discharging control in real-time. The distributed ledger is in charge of recording all the EV charging and discharging data to maintain fairness of power transactions, protects data from being maliciously tampered, and enables the EV user to monitor status of the EV participating in power transactions and dispatching. The intelligent CSMS consists of an artificial intelligence (AI) module, centralized optimal scheduling module, and decentralized optimal control module. The AI module is responsible for forecasting renewable energy generation and load consumption. There is a two-layer architecture consisting of centralized and decentralized optimal control modules; the upper layer performs optimal charging and discharging scheduling of the entire EV charging station at time 15-min time segments, the bottom layer performs distributed optimal scheduling control in each EV charging pile at 5 min time interval. Proposed system in this paper can deal with feeder congestion and real-time power supply and grid demand imbalance, which are caused by high numbers of EVs.

Extending Delivery Range and Decelerating Battery Aging of Logistics UAVs using Public Buses

The battery-powered Unmanned Aerial Vehicle (UAV) is a promising alternative to traditional logistics trucks. Using UAVs can achieve much more speedy, cost-effective, and environment-friendly delivery on an urban scale. However, UAVs suffer from insufficient delivery range and battery aging. This paper presents an innovative logistics UAV scheduling framework using public buses, in which logistics UAVs Land and Recharge its battery on Buses (ULRB) to extend its delivery range and decelerate its fading battery capacity. This work correlates physical layer parameters such as the energy consumption rate, the parcels weight, UAV velocity, the battery temperature to the UAVs path planning, the battery discharging, and the capacity fading models. Specifically, the ULRB framework consists of a single-UAV scheduling module and a multi-UAV dispatching module. In the single-UAV module, a Markov-based algorithm is utilized to plan the UAVs flying path to land and dynamically get recharged on the bus. The latter module optimized the delivery progress in a multi-UAV, multi-parcel, and multi-bus scenario. Finally, using a large-scale real-world bus trajectory dataset, extensive evaluations are conducted to verify ULRB. The results show that ULRB can extend the UAVs delivery range by 5.54 and decelerate the battery aging by 3.26 on average.

Associated outcomes of various iterations of the dedicated orthopaedic trauma room: a literature review

<abstract> <p>Scheduling urgent, orthopaedic trauma cases has long been a challenge for health care institutions. Traditionally, these cases are scheduled for an operating room (OR) slot in the middle of the night, by “bumping” elective cases to later in the day, by adding a case on after-hours, or by delaying the case for several days until an OR becomes available. As a solution to the challenges facing traditional scheduling modules, trauma centers around the country have instituted the use of a dedicated orthopaedic trauma room (DOTR). While there are multiple studies analyzing the effects of DOTRs on various outcomes, there is not a centralized review of these studies. This paper will serve as a review of the various models of the DOTR as well as the effect of the DOTR on after-hours procedures, time to surgery (TTS), duration of surgery (DOS), length of stay (LOS), cost, and surgical complications. An extensive review of the literature was performed through PubMed and Embase. 17 studies were found to meet eligibility criteria. This review suggests that DOTRs have favorable effects on after-hours procedures, cost, and surgical complications. There is variability in the data regarding the effect on TTS, DOS, and LOS.</p> </abstract>

Verification of MILS Partition Scheduling Module Using Layered Methods

MILS partition scheduling module ensures isolation of data between different domains completely by enforcing secure strategies. Although small in size, it involves complicated data structures and algorithms that make monolithic verification of the scheduling module difficult using traditional verification logic (e.g., separation logic). In this paper, we simplify the verification task by dividing data representation and data operation into different layers and then to link them together by composing a series of abstraction layers. The layered method also supports function calls from higher implementation layers into lower abstraction layers, allowing us to ignore implementation details in the lower implementation layers. Using this methodology, we have verified a realistic MILS partition scheduling module that can schedule operating systems (Ubuntu 14.04, VxWorks 6.8 and RTEMS 11.0) located in different domains. The entire verification has been mechanized in the Coq Proof Assistant.

Low area field‐programmable gate array implementation of PRESENT image encryption with key rotation and substitution

Lightweight ciphers are increasingly employed in cryptography because of the high demand for secure data transmission in wireless sensor network, embedded devices, and Internet of Things. The PRESENT algorithm as an ultra-lightweight block cipher provides better solution for secure hardware cryptography with low power consumption and minimum resource. This study generates the key using key rotation and substitution method, which contains key rotation, key switching, and binary-coded decimal-based key generation used in image encryption. The key rotation and substitution-based PRESENT architecture is proposed to increase security level for data stream and randomness in cipher through providing high resistance to attacks. Lookup table is used to design the key scheduling module, thus reducing the area of architecture. Field-programmable gate array (FPGA) performances are evaluated for the proposed and conventional methods. In Virtex 6 device, the proposed key rotation and substitution PRESENT architecture occupied 72 lookup tables, 65 flip flops, and 35 slices which are comparably less to the existing architecture.

Design of a Flexible Exhibition Management Information System for Exhibition Services

In order to solve the problems of low efficiency of management information systems and low utilization rate of the information resources, this research paper proposes the designs of an exhibition management information system by using a B/S structure (Browser/Server mode). The proposed system is aimed at the analysis of functions and types of exhibitions. The system is comprised of the following three components: the browser-side, server-side, and middle layer. Among them, the browser includes a user management module and exhibition environment visualization module. The server includes an upper computer control module and human-computer interaction module. The middle layer includes the data resource management module, document information management module, exhibition project audit approval and management module, and information scheduling module. Experimental results show that the lowest management efficiency of the system is 96.19%, the highest can reach 98.39%, and the utilization rate of information resources is above 94%, indicating that the system has achieved the design expectation.

Guided simulation for dynamic probabilistic risk assessment of complex systems: Concept, method, and application

Probabilistic risk assessment (PRA) is a systematic process of examining how engineered systems work to ensure safety. With the growth of the size of dynamic systems and the complexity of the interactions between hardware, software, and humans, it is extremely difficult to enumerate risky scenarios by the traditional PRA methods. In this study, a new dynamic probabilistic risk assessment methodology is proposed that employs a new exploration strategy to generate risky scenarios. The proposed methodology consists of three main modules, including simulation, planner, and scheduler. In this methodology, the engineering knowledge of the system is explicitly used to guide the simulation module to achieve higher efficiency and accuracy. The engineering knowledge is reflected in the planner module which is responsible for generating plans as a high-level map to guide the simulation. The scheduler module is responsible for guiding the simulation by controlling the timing and occurrence of the random events. In this paper, modules of the proposed methodology, and their interactions are explained in detail. The developed methodology is used to perform risk assessment of a Space Shuttle ascent phase, and results show the effectiveness of the proposed platform.

Energy management system for microgrids using weighted salp swarm algorithm and hybrid forecasting approach

The concept of a microgrid (MG) has been introduced to integrate the conventional generators, different renewable energy resources and energy storage systems (ESS) to meet the specific load demand. However, the intermittent nature of renewable energy resources produces a variable output, which drives an imbalance between power generation and demand in MG. The ESS is utilized to makes a balance between power generation and demand. When several renewable energy resources and ESS are available in MG as energy resources, then an energy management system (EMS) is required that can handle the stochastic nature of renewable energy resources, schedule the power of renewable energy resources and ESS for managing the power flow among MG resources and main grid while ensuing cost-effective operation. Therefore, this paper proposed an optimum EMS that aims to minimize the overall operating cost of grid-connected MG along with the short-term forecasting of PV power and load demand. The proposed EMS consists of four modules: forecasting, scheduling, data acquisition (DAQ), and human–machine interface (HMI) modules. An improved hybrid forecasting approach that combines a 3-level stationary wavelet transform (SWT) and grey wolf optimization-based least-square support vector machine (GWO-LSSVM) is proposed in the forecasting module to achieve day-ahead forecasting of PV power and load demand. In the scheduling module, the weighted salp swarm algorithm-based scheduling is applied to achieve the optimum power flow of grid-connected MG. Then, the DAQ and HMI module is used to monitor, analyze, and modified the input variables of the forecasting and scheduling module. The MATLAB/Simulink environment is then used to simulate the proposed EMS for grid-connected MG. Finally, numerical results demonstrate the efficiency of the proposed EMS for grid-connected MG with commercial load demand over the existing competitive approaches.

Predictive maintenance integrated production scheduling by applying deep generative prognostics models: approach, formulation and solution

To harness the full potential of predictive maintenance (PdM), PdM information has to be used to optimally plan production and maintenance actions. Hence, operation-specific modelling of degradation, i.e. predictions of the health condition under time-varying operational conditions, has to be realized. By utilizing operation-specific degradation information, maintenance and production can be planned with regard to each other and thus, predictive maintenance integrated production scheduling (PdM-IPS) is enabled. This publication proposes a novel PdM-IPS approach consisting of two interacting modules: an operation-specific Prognostics and Health Management (PHM) module and an integrated production scheduling and maintenance planning (IPSMP) module. Specifically, the mathematical problem of the IPSMP module based on an extended version of the maintenance integrated flexible job shop problem is formulated. A two-stage genetic algorithm to efficiently solve this problem is designed and subsequently applied to simulated condition monitoring, as well as real industrial data. Results indicate that the approach is able to find feasible high quality PdM integrated production schedules.

Problematika Pelayanan Bimbingan Manasik Haji Sepanjang Tahun di Pusat Layanan Haji dan Umrah Terpadu kota Padang

This research describes information on various issues of Hajj Manasik Guidance activities throughout the year at the Integrated Hajj and Umrah Service Center (PLHUT) Padang City. This activity is an innovation from the Ministry of Religion of the Republic of Indonesia since 2019 which aims to meet the achievement of the objectives of organizing the Hajj, providing guidance, service and protection for pilgrims in accordance with sharia, as well as realizing independence and resilience in the implementation of the Hajj and Umrah pilgrimages. The Facts show that this new innovation creates some problems for PLHUT that require solutions. The research aims to find out the problems and solutions provided in dealing with the problem of guidance. This research is a field research with a descriptive qualitative approach. The data source is taken from the Head of the Hajj and Umrah Organizing Section of the Ministry of Religion of the city of Padang, implementing officers, supervisors and prospective pilgrims as well, by using purposive technique. The instruments that used are observation and interview. The results are as following : 1) The problems faced by PLHUT were budget issues, guidance materials, Jamaat do not have the maximum time to follow the guidance, the available time allocation and limited facilities. Presentation of guidance material and its schedule are less systematic and module are ineffective. 2) PLHUT provides a solution by increasing the guidance time in “Ramadan” every day, the committee will provide a schedule and guidance module.

Optimized resource-constrained method for project schedule compression

PurposeThis paper aims to present an integrated method for optimized project duration and costs, considering the size and cost of crews assigned to project activities' execution modes.Design/methodology/approachThe proposed method utilizes fuzzy set theory (FSs) for modeling uncertainties associated with activities' duration and cost and genetic algorithm (GA) for optimizing project schedule. The method has four main modules that support two optimization methods: modeling uncertainty and defuzzification module; scheduling module; cost calculations module; and decision-support module. The first optimization method uses the elitist non-dominated sorting genetic algorithm (NSGA-II), while the second uses a dynamic weighted optimization genetic algorithm. The developed scheduling and optimization methods are coded in python as a stand-alone automated computerized tool to facilitate the developed method's application.FindingsThe developed method is applied to a numerical example to demonstrate its use and illustrate its capabilities. The method was validated using a multi-layered comparative analysis that involves performance evaluation, statistical comparisons and stability evaluation. Results indicated that NSGA-II outperformed the weighted optimization method, resulting in a better global optimum solution, which avoided local minima entrapment. Moreover, the developed method was constructed under a deterministic scenario to evaluate its performance in finding optimal solutions against the previously developed literature methods. Results showed the developed method's superiority in finding a better optimal set of solutions in a reasonable processing time.Originality/valueThe novelty of the proposed method lies in its capacity to consider resource planning and project scheduling under uncertainty simultaneously while accounting for activity splitting.

DESIDS: An Integrated Decision Support System for the Planning, Analysis, Management and Rehabilitation of Pressurised Irrigation Distribution Systems

Pressurized irrigation distribution systems (PIDSs) play a vital role in irrigation intensification, especially in the Mediterranean region. The design, operation and management of these systems can be complex, as they involve several intertwined processes which need to be considered simultaneously. For this reason, numerous decision support systems (DSSs) have been developed and are available to deal with these processes, but as independent components. To this end, a comprehensive DSS called DESIDS has been developed and tested. This DSS has been developed to bear in mind the needs of irrigation district managers for an integrated tool that can assist them in taking strategic decisions for managing and developing reliable, adequate and sustainable water distribution plans which provide the best services to farmers. Hence, four modules were integrated in DESIDS: (i) irrigation demand and scheduling module; (ii) hydraulic analysis module; (iii) operation and management module; and (iv) design and rehabilitation module. DESIDS was tested on different case studies, proving itself a valuable tool for irrigation district managers, as it provides a wide range of decision options for the proper operation and management of PIDSs. The developed DSS can be used as a platform for future integrations and expansions, and to include other processes needed for better decision-making support.

A* Based Routing and Scheduling Modules for Multiple AGVs in an Industrial Scenario

A multi-AGV based logistic system is typically associated with two fundamental problems, critical for its overall performance: the AGV’s route planning for collision and deadlock avoidance; and the task scheduling to determine which vehicle should transport which load. Several heuristic functions can be used according to the application. This paper proposes a time-based algorithm to dynamically control a fleet of Autonomous Guided Vehicles (AGVs) in an automatic warehouse scenario. Our approach includes a routing algorithm based on the A* heuristic search (TEA*—Time Enhanced A*) to generate free-collisions paths and a scheduling module to improve the results of the routing algorithm. These modules work cooperatively to provide an efficient task execution time considering as basis the routing algorithm information. Simulation experiments are presented using a typical industrial layout for 10 and 20 AGVs. Moreover, a comparison with an alternative approach from the state-of-the-art is also presented.

HGP4CNN: an efficient parallelization framework for training convolutional neural networks on modern GPUs

Graphics Processing Units (GPUs) have evolved into powerful accelerators for the development of Convolutional Neural Network (CNN) models. Most existing GPU-based frameworks adopt a kernel-based execution approach and only focus on optimizing individual kernels for better performance and resource utilization. With this approach, kernels involved will be launched sequentially, which may result in the underutilization of GPU resources due to the limited optimization space of a single kernel. In this paper, we propose an efficient software parallelization framework, called HGP4CNN, to accelerate the training of CNN models by considering the characteristics of workloads from both the same layer and adjacent layers as well as new GPU features such as concurrent kernel execution. In the intra-layer level, to achieve a better training performance of a single network layer, we design a novel model-based lightweight parallelization module to make better use of the concurrent kernel execution feature on modern GPUs. An asynchronous resource tracker is used to collect kernels’ information at runtime, and a kernel analyzer is devised to calculate the number of kernels that can be dispatched concurrently. Moreover, to avoid consuming too many CPU threads or process resources, we integrate a runtime scheduler module for kernel launch and a pool-based stream manager for GPU work queue management. While in the inter-layer level, we present a pipeline execution strategy to overlap the processing of workloads from adjacent layers. To determine the number of samples to be processed by a single pipeline stage, the analysis result from the intra-layer module is considered. In the end, we implement a prototype of the proposed framework with Caffe, a well-known deep learning framework and conduct experiments with four off-the-shelf CNN models on three NVIDIA GPUs. Results show that HGP4CNN can be exploited to achieve better performance over the original implementation and keep the convergence property of networks. We can achieve a speedup of up to 6.X for a single convolutional layer and 2.X for multiple layers within pipelines of a network model.

Mass Tourism Data Analysis API Based on E-Commerce Platform

With the rapid development of Internet information technology, e-commerce transactions are accepted and known by more and more people, and more and more attention is paid by major enterprises. This paper mainly studies the design and implementation of mass tourism data analysis API based on e-commerce platform. The software system of this paper adopts the method of black box testing, which basically covers the main business logic of each functional module. The class and method of task scheduling, monitoring alarm and API gateway service module are tested, which basically covers the logic of class method of these modules. The system performance test is mainly to test the system by recording test script by automatic test tool. In this paper, the system response speed, concurrency and stability index are tested by using the automatic test tool load runner. The data shows that when the number of API requests reaches 4200, there will be a certain number of requests failed to respond to processing. The results show that the system has achieved the design goals in advance and basically completed the requirements listed in the requirements.

Parameterization and modeling of paddy rice (Oryza sativa L. ssp. japonica) growth and water use in cold regions: Yield and water-saving analysis

Rice is the high-water consumption crop that feeds more than half of the world population. The planting area of cold region rice is expanded rapidly in recent years due to the high market values, also resulting in increasing water shortages and eco-environmental problems on a regional scale. Pursuing high benefits and increasing water use efficiency become more significant both for stakeholders and managers. To understand the yield variation and water-saving potential, the specific rice model-ORYZA2000 was adopted to predict the responses of rice growth/yield and water use to the various climatic and irrigation conditions in a typical cold region of northeast China. Field experiments were conducted for the Japonica rice in cold regions (JRC) during 2018 and 2019, with detailed observations on rice growth and development. The ORYZA2000 was calibrated and validated using two-year experimental data. Simulated ponded water depth, various organ development and yield fitted well with field observations. A set of recommended values of crop parameters for JRC was obtained through the systematic parametrization, showing very specific and different characters from the varieties in warmer regions. The irrigation scheduling module was modified for adapting to the various irrigation regimes. After that, the scenario simulation with 30-year different climatic data (i.e. 1990–2019) and three different irrigation regimes (i.e. traditional flood irrigation (TFI), shallow-wet irrigation (SWI) and controlled irrigation (CTI)) were conducted using the calibrated model. Status evaluation and scenario analyses indicated that the crop yield of JRC varied significantly due to relatively large fluctuations of photothermal conditions in different years. The inter-annual potential yield varied from 6790 to 9516 kg ha−1, with an average of 8210 kg ha−1. Evaluation on water-saving scenarios indicated that adopting SWI and CTI regimes could averagely save up to 90 mm and 140 mm (20% and 30%) of irrigation water in dry years, respectively, compared with TFI regime. Overall, the obtained set of parameters for JRC could efficiently complement the parameter database for rice modeling in cold regions; meanwhile, the yield and water-saving evaluation could provide very useful information and the basis for supporting the future JRC production and water-saving management.

Determination of Grain Yield Inputs of the Maize Hybrid Giza 168 Using a Six-Factor Central Composite Design in Mediterranean Regions Under Irrigation

The maize single hybrid Giza 168 was evaluated for grain yield in a six-factor central composite design in 2015 and 2016 summer seasons. The six factors included; surface irrigation level (I), potassium (K), phosphorus (P), nitrogen (N), sowing date (SD), and plant density (PD). The CROPWAT schedule module was applied for evaluation of irrigation practices and to develop alternative improved water delivery schedules. The six studied factors explained 57% of the variation in grain yield, with significant linear effects for PD (0.48*), I×K interaction (-0.45*), and P×N interaction (-0.69*). The highest grain yields (8.05 and 8.06 t/ha) were obtained from two combinations, i.e., high irrigation with high K, low P, high N, late SD, and high PD; and low irrigation with high K, high P, low N, late SD, and low PD, respectively. Simulation of irrigation scheduling indicated that the quantity of irrigation water could be reduced by 22% and the irrigation interval extended to 15 days without any loss in grain yield. The results showed that application of 270 kg N/ha, 100 kg K2O/ha, 34.5 kg P2O5/ha, and sowing at 20 to 30 May with a plant density of 65,000 plants/ha will realize the highest yield potential of the hybrid Giza 168.