Cloud computing datacenters consume hundreds of KWs of power per hour, thus promoting the need for optimizing the energy consumption of datacenters. Job scheduling is at the heart of any successful power management technique used in datacenters. Power consumption of computing nodes in datacenters is mainly determined by the CPU, memory, disk storage and the network interfaces, with the CPU consuming the major portion. Therefore, in this work we focus on the efficient utilization of datacenter CPUs to optimize energy consumption. The scheduling problem is modeled using Integer Linear Programming (ILP) techniques, where a model is formulated with the objective of minimizing the total power consumed by the active and idle cores of the CPUs while meeting a set of constraints. This work focuses on assessing the performance of state-of-the art generic ILP and 0-1 SAT-based ILP solvers in solving ILP formulations of the scheduling problem.
Dynamic nuclear polarization (DNP) effect is used to enhance the nuclear magnetic resonance (NMR) by adding paramagnetic polarizing agents (containing unpair electrons) to the tested sample. Those polarizing agents are usually organic materials with a limited number of unpair electrons. Semiconductor materials along with spintronic devices provide alternative and more efficient substitute for the polarizing agents. Since, the number of spin polarized electrons can be controlled by changing doping profile and biasing conditions. We discussed the DNP effect in semiconductor materials along with spintronic devices.
With the recent advances in internet of vehicles (IoV), the traditional traffic control problem at intersections is being revisited to cope with the future intelligent transportation systems. In this paper, we introduce the performance evaluation of a distributed IoV-based intersection traffic control protocol. Extensive simulations have been conducted using various real traffic scenarios to evaluate the efficiency and the dynamics of the newly proposed traffic control approach that gives the intelligent vehicles full autonomous control of crossing intersections using vehicle-to-vehicle and vehicle-to-infrastructure communications. Simulation results using the tools OMNet++, VEINS, and SUMO showed that the new approach is a viable alternative and addition to the existing traffic control approaches.
Wireless communications makes the delivery of real time information at hands. This includes information from Vehicle to Infrastructure (V2I) and Vehicle to Vehicle (V2V) communications. In this paper, the focus will be on the flow of traffic across number of adjacent Road Side Units (RSU) in various intersections. The objective is to allow RSUs' (I2I) cooperation by exchanging information that is originally collected from vehicles through (V2I). Advanced knowledge of the moving vehicles will lead to better traffic management at intersections and will reduce waiting time
Simulations on Vehicular Ad Hoc Network (VANET) relies on many parameters such as speed, driving skills, lane changing, etc. It is desirable to reproduce such conditions in a simulation environment. Recently, we have seen lots of developments in this area for example projects like OMNet++, Network Simulator 3, OPNET, and others. All these projects require a way to import real data from map sources to make the model more realistic. A road or highway with its main conditions is today feasible with many collaborators in projects like OpenStreetMap which offers map data for research which must be imported into a simulation scenario. We propose a procedure to achieve this with SUMO (Simulation of Urban Mobility) so that it generates a network mobility file that could be used for further studies in VANET simulations. In this procedure, a quantity of vehicles and inherent features of exported maps are considered.
Project management is critical for companies to stay competitive which make material and human resources management an issue of increasing importance. The project management scheduling process of deciding when an activity start and how resources will be used will highly impact the project duration and cost. Resource-constrained project scheduling problem (RCPSP) considers activities of known durations linked by precedence relations and resource requests from resources of limited availability to find a schedule of minimal duration. Traditionally, the RCPSP considered the objective of makespan minimization to plan the overall project; however, since the value of money decreases with time, it is critical to incorporate the financial aspect of the project and schedule the activities in such a way that will maximize the profit. In this work, we will propose a mathematical model for the multi-mode resource-constrained project scheduling problem with material ordering to maximize the net present value (MMRCSPMODC). The problem will be subjected to precedence constrains, project completion constraints and materials constraints as well as penalties that will be imposed in the case of any delays. Project scheduling and material ordering decisions will be emphasized to determine the time and quantity of an order since setting the material ordering decisions after the project scheduling phase leads to non-optimal solutions. In addition to the developing the mathematical model, we will propose a heuristic approach to obtain near optimal solution since the problem under study is NP-hard.
This work considers a design model for a Fractional Slot Concentrated Winding (FSCW) Permanent Magnet Synchronous Machine (PMSM) for an Electric Vehicle (EV) to obtain a minimal loss and machine mass. A population-based multi- objective optimization design is utilized to design and determine the machine parameters. The results of the optimization show there is an inverse relation between the total loss and total mass of the machine. Validation of the results is achieved by means of two dimensional Finite Element Analysis (FEA).
The goal of this work is to present a suitable hardware realization for an artificial neuron with a sigmoid activation function and fixed-point data representation. Therefore, this paper presents all the fixed-point arithmetical operations needed to realize a neuron with a sigmoid activation function. In addition, the paper summarizes the different hardware implementations of the required operations on Field Programmable Gate Array (FPGA). All in all, the paper provides an analysis of the different implementation possibilities.
In this paper, Load Frequency Control (LFC) is designed to a single-area power grid connected to a photovoltaic (PV) system. Both the power grid and the PV systems have been modeled separately. Three controllers have been designed for this connected system: Linear Quadratic Regulator (LQR), PI, and Fuzzy Logic Controller (FLC) to regulate the error of frequency deviation. A comparison has been carried out between the three methods to check the best performance in terms of settling time, undershoot and steady state error. The criteria to be met in the power plant according to UAE standards are settling time less than 3s, undershoot less than 0.02 Hz and a steady state error of 0. LQR met all three criteria for the system under study, FLC improved the system response greatly and is useful for systems with complex models, while the PI did not meet two of the specifications required.
This article elaborates the novel approach for air-gap magnetic field analysis of switched reluctance motor in case of rotor cracks and rotor tilt around its shaft axis. The fault diagnosis is illustrated on the basis of a 3-D model of the SRM using of a technique known as finite element analysis (FEA). The flux linkage analytical equations are used for the derivation of inductance expressions. The results obtained from the 3-D FEA of a SRM having 6 stator and 4 rotor poles shows the variation of mutual inductance with the cracked rotor conditions and the tilting of rotor shaft. The results obtained explain the usefulness for the detection of cracked rotors and shaft tilting.
