Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “distributed scheduling”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
NASA NTRS2002

A De-centralized Scheduling and Load Balancing Algorithm for Heterogeneous Grid Environments

In the past two decades, numerous scheduling and load balancing techniques have been proposed for locally distributed multiprocessor systems. However, they all suffer from significant deficiencies when extended to a Grid environment: some use a centralized approach that renders the algorithm unscalable, while others assume the overhead involved in searching for appropriate resources to be negligible. Furthermore, classical scheduling algorithms do not consider a Grid node to be N-resource rich and merely work towards maximizing the utilization of one of the resources. In this paper, we propose a new scheduling and load balancing algorithm for a generalized Grid model of N-resource nodes that not only takes into account the node and network heterogeneity, but also considers the overhead involved in coordinating among the nodes. Our algorithm is decentralized, scalable, and overlaps the node coordination time with that of the actual processing of ready jobs, thus saving valuable clock cycles needed for making decisions. The proposed algorithm is studied by conducting simulations using the Message Passing Interface (MPI) paradigm.

Arora, Manish↗
NASA NTRS2002

A De-Centralized Scheduling and Load Balancing Algorithm for Heterogeneous Grid Environments

In the past two decades, numerous scheduling and load balancing techniques have been proposed for locally distributed multiprocessor systems. However, they all suffer from significant deficiencies when extended to a Grid environment: some use a centralized approach that renders the algorithm unscalable, while others assume the overhead involved in searching for appropriate resources to be negligible. Furthermore, classical scheduling algorithms do not consider a Grid node to be N-resource rich and merely work towards maximizing the utilization of one of the resources. In this paper we propose a new scheduling and load balancing algorithm for a generalized Grid model of N-resource nodes that not only takes into account the node and network heterogeneity, but also considers the overhead involved in coordinating among the nodes. Our algorithm is de-centralized, scalable, and overlaps the node coordination time of the actual processing of ready jobs, thus saving valuable clock cycles needed for making decisions. The proposed algorithm is studied by conducting simulations using the Message Passing Interface (MPI) paradigm.

Arora, Manish↗
NASA NTRS1991

Architectural impact of FDDI network on scheduling hard real-time traffic

The architectural impact on guaranteeing synchronous message deadlines in FDDI (Fiber Distributed Data Interface) token ring networks is examined. The FDDI network does not have facility to support (global) priority arbitration which is a useful facility for scheduling hard real time activities. As a result, it was found that the worst case utilization of synchronous traffic in an FDDI network can be far less than that in a centralized single processor system. Nevertheless, it is proposed and analyzed that a scheduling method can guarantee deadlines of synchronous messages having traffic utilization up to 33 pct., the highest to date.

Agrawal, Gopal↗
NASA NTRS2004

Job Scheduling in a Heterogeneous Grid Environment

Computational grids have the potential for solving large-scale scientific problems using heterogeneous and geographically distributed resources. However, a number of major technical hurdles must be overcome before this potential can be realized. One problem that is critical to effective utilization of computational grids is the efficient scheduling of jobs. This work addresses this problem by describing and evaluating a grid scheduling architecture and three job migration algorithms. The architecture is scalable and does not assume control of local site resources. The job migration policies use the availability and performance of computer systems, the network bandwidth available between systems, and the volume of input and output data associated with each job. An extensive performance comparison is presented using real workloads from leading computational centers. The results, based on several key metrics, demonstrate that the performance of our distributed migration algorithms is significantly greater than that of a local scheduling framework and comparable to a non-scalable global scheduling approach.

Shan, Hong-Zhang↗
DOE OSTI2020

Modular performance prediction for scientific workflows using Machine Learning

Scientific workflows provide an opportunity for declarative computational experiment design in an intuitive and efficient way. A distributed workflow is typically executed on a variety of resources, and it uses a variety of computational algorithms or tools to achieve the desired outcomes. Such a variety imposes additional complexity in scheduling these workflows on large scale computers. As computation becomes more distributed, insights into expected workload that a workflow presents become critical for effective resource allocation. In this paper, we present a modular framework that leverages Machine Learning for creating precise performance predictions of a workflow. The central idea is to partition a workflow in such a way that makes the task of forecasting each atomic unit manageable and gives us a way to combine the individual predictions efficiently. We recognize a combination of an executable and a specific physical resource as a single module. This gives us a handle to characterize workload and machine power as a single unit of prediction. Overall, our modular technique of creating atomic modules and deployment of longest-path approach to estimate workflow performance, allows the framework to adapt to highly complex nested directed acyclic workflows and scale to new scenarios, since it does not make assumptions of underlying workflow structure. We present performance estimation results of independent workflow modules executed on the XSEDE SDSC Comet cluster using various Machine Learning algorithms. The results provide insights into the behavior and effectiveness of different algorithms in the context of scientific workflow performance prediction.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Deep Reinforcement Scheduling of Energy Storage Systems for Real-time Voltage Regulation in Unbalanced LV Networks with High PV Penetration

The ever-growing higher penetration of distributed energy resources (DERs) in low-voltage (LV) distribution systems brings both opportunities and challenges to voltage support and regulation. This paper proposes a deep reinforcement learning (DRL)-based scheduling scheme of energy storage systems (ESSs) to mitigate system voltage deviations in unbalanced LV distribution networks. The ESS-based voltage regulation problem is formulated as a multi-stage quadratic stochastic program, with the objective of minimizing the expected total daily voltage regulation cost while satisfying operational constraints. While existing voltage regulation methods are mostly focused on onetime- step control, this paper explores a day-horizon systemwide voltage regulation problem. In other words, the size of action and state spaces are extremely high-dimensional and need to be delicately handled. Furthermore, in order to overcome the difficulty of modeling uncertainties and develop a realtime solution, a learn-to-schedule feedback control framework is proposed by adapting the problem to a model-free DRL setting. The proposed algorithm is tested on a customized 6-bus system and a modified IEEE 34-bus system. Simulation results validate the effectiveness and near-optimality of voltage regulation by ESS in comparison with a deterministic quadratic program solution.

Wang, Shengyi↗
NASA NTRS2005

Effect of Sampling Schedule on Pharmacokinetic Parameter Estimates of Promethazine in Astronauts

Six astronauts on the Shuttle Transport System (STS) participated in an investigation on the pharmacokinetics of promethazine (PMZ), a medication used for the treatment of space motion sickness (SMS) during flight. Each crewmember completed the protocol once during flight and repeated thirty days after returned to Earth. Saliva samples were collected at scheduled times for 72 h after PMZ administration; more frequent samples were collected on the ground than during flight owing to schedule constraints in flight. PMZ concentrations in saliva were determined by a liquid chromatographic/mass spectrometric (LC-MS) assay and pharmacokinetic parameters (PKPs) were calculated using actual flight and ground-based data sets and using time-matched sampling schedule on ground to that during flight. Volume of distribution (V(sub c)) and clearance (Cl(sub s),) decreased during flight compared to that from time-matched ground data set; however, Cl(sub s) and V(sub c) estimates were higher for all subjects when partial ground data sets were used for analysis. Area under the curve (AUC) normalized with administered dose was similar in flight and partial ground data; however AUC was significantly lower using time-matched sampling compared with the full data set on ground. Half life (t(sub 1/2)) was longest during flight, shorter with matched-sampling schedule on ground and shortest when complete data set from ground was used. Maximum concentration (C(sub max)), time for C(sub max), (t(sub max)), parameters of drug absorption, depicted a similar trend with lowest and longest respectively, during flight, lower with time-matched ground data and highest and shortest with full ground data.

Boyd, Jason L.↗
NASA NTRS1987

Comparing the Performance of Two Dynamic Load Distribution Methods

Parallel processing of symbolic computations on a message-passing multi-processor presents one challenge: To effectively utilize the available processors, the load must be distributed uniformly to all the processors. However, the structure of these computations cannot be predicted in advance. go, static scheduling methods are not applicable. In this paper, we compare the performance of two dynamic, distributed load balancing methods with extensive simulation studies. The two schemes are: the Contracting Within a Neighborhood (CWN) scheme proposed by us, and the Gradient Model proposed by Lin and Keller. We conclude that although simpler, the CWN is significantly more effective at distributing the work than the Gradient model.

Kale, L. V.↗
DOE OSTI2022

Simulation-based assessment on stochastic load scheduling for building cooling systems

Here, to fill knowledge gaps related to stochastic load scheduling, we performed a comprehensive evaluation of the stochastic load scheduling for building cooling systems. Specifically, we studied the common uncertain variables in the load scheduling process for building cooling systems and categorized those variables based on their dynamic patterns. We then developed a generic stochastic load scheduling framework and applied it to building cooling systems that served a simulated community. This community consists of 100 heterogeneous houses and serves as a virtual testbed for evaluating the performance of stochastic load scheduling. In this evaluation, we considered representatives of uncertain variables with different dynamic patterns and included 100 realizations of the considered uncertainty in the evaluation to better catch the probability distribution of the control performance. The evaluation results suggest that deterministic load scheduling can reduce the operating energy cost by 18% but its performance can be affected by uncertainty. Stochastic load scheduling can further decrease the operating energy cost under uncertainty compared to deterministic load scheduling. We also found that the effectiveness of stochastic load scheduling in handling uncertainty is not directly associated with the number of uncertainty scenarios that are considered in its formulation.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
NASA NTRS1989

Run-time scheduling and execution of loops on message passing machines

Sparse system solvers and general purpose codes for solving partial differential equations are examples of the many types of problems whose irregularity can result in poor performance on distributed memory machines. Often, the data structures used in these problems are very flexible. Crucial details concerning loop dependences are encoded in these structures rather than being explicitly represented in the program. Good methods for parallelizing and partitioning these types of problems require assignment of computations in rather arbitrary ways. Naive implementations of programs on distributed memory machines requiring general loop partitions can be extremely inefficient. Instead, the scheduling mechanism needs to capture the data reference patterns of the loops in order to partition the problem. First, the indices assigned to each processor must be locally numbered. Next, it is necessary to precompute what information is needed by each processor at various points in the computation. The precomputed information is then used to generate an execution template designed to carry out the computation, communication, and partitioning of data, in an optimized manner. The design is presented for a general preprocessor and schedule executer, the structures of which do not vary, even though the details of the computation and of the type of information are problem dependent.

Crowley, Kay↗
NASA NTRS1990

Run-time scheduling and execution of loops on message passing machines

Sparse system solvers and general purpose codes for solving partial differential equations are examples of the many types of problems whose irregularity can result in poor performance on distributed memory machines. Often, the data structures used in these problems are very flexible. Crucial details concerning loop dependences are encoded in these structures rather than being explicitly represented in the program. Good methods for parallelizing and partitioning these types of problems require assignment of computations in rather arbitrary ways. Naive implementations of programs on distributed memory machines requiring general loop partitions can be extremely inefficient. Instead, the scheduling mechanism needs to capture the data reference patterns of the loops in order to partition the problem. First, the indices assigned to each processor must be locally numbered. Next, it is necessary to precompute what information is needed by each processor at various points in the computation. The precomputed information is then used to generate an execution template designed to carry out the computation, communication, and partitioning of data, in an optimized manner. The design is presented for a general preprocessor and schedule executer, the structures of which do not vary, even though the details of the computation and of the type of information are problem dependent.

Saltz, Joel↗
DOE OSTI2021

Enhancing Active Distribution Systems Resilience by Fully Distributed Self-Healing Strategy

Distributed restoration can exploit smart grid technologies to enhance the resilience of active distribution networks toward a self-healing smart grid. However, the large number of decision variables, especially the binary ones for reconfiguration, bring challenges to developing scalable distributed distribution service restoration (DDSR) strategies. This paper proposes a fully distributed solution procedure based on the alternating direction method of multipliers (ADMM) for mixed-integer programming problems and applies to develop the DDSR framework. The method consists of relax-drive-polish phases, 1) relaxing binary variables, and applying the convex ADMM as a warm start; 2) driving the solutions toward Boolean values through a proximal operator; 3) fixing the obtained binding binary variables and solving the rest of the problem to polish results and achieve a high-quality suboptimal solution. Then, an autonomous clustering strategy and consensus ADMM are integrated with the proposed method to realize the fully distributed cluster-based framework of DDSR. This framework can first determine DER scheduling and switch status for reconfiguration to energize the out-of-service areas from local faults, and then provide the load restoration solution in a distributed manner for total blackouts in large-scale distribution networks. Furthermore, the effectiveness and scalability of the proposed DDSR framework are demonstrated through testing on the IEEE 123-node, IEEE 8500-node, and synthetic 100k-node test feeders.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2020

Approximating Trajectory Constraints With Machine Learning – Microgrid Islanding With Frequency Constraints

Here, we introduce deep earning aided constraint encoding to tackle the frequency-constraint microgrid scheduling problem. The nonlinear function between system operating condition and frequency nadir is approximated by using a neural network, which admits an exact mixed-integer formulation (MIP). This formulation is then integrated with the scheduling problem to encode the frequency constraint. With the stronger representation power of the neural network, the resulting commands can ensure adequate frequency response in a realistic setting in addition to islanding success. The proposed method is validated on a modified 33-node system. Successful islanding with a secure response is simulated under the scheduled commands using a detailed three-phase model in Simulink. The advantages of our model are particularly remarkable when the inertia emulation functions from wind turbine generators are considered.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRSDate not supplied

Design of Safe Separation Bounds for Temporally Deconflicted Trajectories Under Bounded Uncertainties

This paper explores the derivation of safe separation bounds for a heterogeneous group of~$n$ Uncrewed Aerial Systems (UAS) that are assigned temporally deconflicted trajectories. Compared to spatially deconflicted trajectories, temporal deconfliction can lead to higher traffic capacities and a more efficient use of the available airspace. One challenge with this type of deconfliction is that collisions can occur if some cooperating UAS are behind or ahead of schedule. To overcome this risk, this paper derives a lower bound on the safety distance between two heterogeneous UAS in the presence of bounded uncertainties. This safety distance can be leveraged to inform trajectory generation algorithms. The proposed bound establishes a rigorous safety margin when the fleet deviates from the planned trajectories, both temporally and spatially. For its derivation the paper assumes the UAS implement a distributed coordination algorithm that allows the fleet to maintain their schedules synchronized within a bounded temporal error, and a path-following algorithm that lets the vehicles track a target that moves along the planned trajectory with a bounded spatial error.

autonomy↗
NASA NTRS2003

Improving Resource Selection and Scheduling Using Predictions

The introduction of computational grids has resulted in several new problems in the area of scheduling that can be addressed using predictions. The first problem is selecting where to run an application on the many resources available in a grid. Our approach to help address this problem is to provide predictions of when an application would start to execute if submitted to specific scheduled computer systems. The second problem is gaining simultaneous access to multiple computer systems so that distributed applications can be executed. We help address this problem by investigating how to support advance reservations in local scheduling systems. Our approaches to both of these problems are based on predictions for the execution time of applications on space- shared parallel computers. As a side effect of this work, we also discuss how predictions of application run times can be used to improve scheduling performance.

Smith, Warren↗
NASA NTRS1994

Coordinated scheduling for dynamic real-time systems

In this project, we addressed issues in coordinated scheduling for dynamic real-time systems. In particular, we concentrated on design and implementation of a new distributed real-time system called R-Shell. The design objective of R-Shell is to provide computing support for space programs that have large, complex, fault-tolerant distributed real-time applications. In R-shell, the approach is based on the concept of scheduling agents, which reside in the application run-time environment, and are customized to provide just those resource management functions which are needed by the specific application. With this approach, we avoid the need for a sophisticated OS which provides a variety of generalized functionality, while still not burdening application programmers with heavy responsibility for resource management. In this report, we discuss the R-Shell approach, summarize the achievement of the project, and describe a preliminary prototype of R-Shell system.

Natarajan, Swaminathan↗
NASA NTRS2000

Creep Properties of NiAl-1Hf Single Crystals Re-Investigated

NiAl-1Hf single crystals have been shown to be quite strong at 1027 C, with strength levels approaching those of advanced Ni-based superalloys. Initial testing, however, indicated that the properties might not be reproducible. Study of the 1027 C creep behavior of four different NiAl-1Hf single-crystal ingots subjected to several different heat treatments indicated that strength lies in a narrow band. Thus, we concluded that the mechanical properties are reproducible. Recent investigations of the intermetallic NiAl have confirmed that minor alloying additions combined with single-crystal growth technology can produce elevated temperature strength levels approaching those of Ni-based superalloys. For example, General Electric alloy AFN 12 {Ni-48.5(at.%) Al-0.5Hf-1Ti-0.05Ga} has a creep rupture strength equivalent to Rene 80 combined with a approximately 30-percent lower density, a fourfold improvement in thermal conductivity, and the ability to form a self-protective alumina scale in aggressive environments. Although the compositions of strong NiAl single crystals are relatively simple, the microstructures are complex and vary with the heat treatment and with small ingot-toingot variations in the alloy chemistry. In addition, initial testing suggested a strong dependence between microstructure and creep strength. If these observations were true, the ability to utilize NiAl single-crystal rotating components in turbine machinery could be severely limited. To investigate the possible limitations in the creep response of high-strength NiAl single crystals, the NASA Glenn Research Center at Lewis Field initiated an in depth investigation of the effect of heat treatment on the microstructure and subsequent 1027 C creep behavior of [001]-oriented NiAl-1Hf with a nominal chemistry of Ni-47.5Al-1Hf-0.5Si. This alloy was selected since four ingots, grown over a number of years and possessing slightly different compositions, were available for study. Specimens taken from the ingots were subjected to several heat treatment schedules, examined by transmission electron microscopy, and tested in both compression and tension. An example of the microstructure found in a [001]-oriented NiAl-1Hf specimen after a solution treatment at 1317 C for 50 hr followed by air cooling is illustrated in the image on the left, where the NiAl matrix contains a uniform distribution of nanometer-scale Gphase (Ni16Hf6Si7) precipitates. Other heat treating schedules produced microstructures with nanometer-sized G-phase cubes and plates or, in an extreme case, produced a microstructure with all the G-phase converted to Heusler (Ni2AlHf) particles. The results of 1027 C creep strength and strain rate testing are illustrated which summarizes data from tensile and compressive testing of samples cut from all four NiAl-1Hf ingots and subjected to a variety of heat treatment schedules. With one exception, all the strength values lie in a narrow band that spans six orders of magnitude in strain rate. The only factor that produced results outside of this band was the heat treatment schedule that dissolved all the G-phase and replaced it with Heusler precipitates. The results portrayed in this figure lead to the important practical conclusion that the elevated-temperature creep properties of NiAl-1Hf single crystals are reproducible and are not affected by small variations in alloy chemistry from ingot to ingot or by different initial distributions of G-phase in the heat-treated alloy. The only variable in this study that produced a significant and delerious effect on mechanical strength was a post-solution heat treatment that lead to the complete disappearance of the G-phase in favor of Heusler precipitates.

Whittenberger, J. Daniel↗