Job Completion Rate Research Articles

A Deep Reinforcement Learning Method Based on a Transformer Model for the Flexible Job Shop Scheduling Problem

The flexible job shop scheduling problem (FJSSP), which can significantly enhance production efficiency, is a mathematical optimization problem widely applied in modern manufacturing industries. However, due to its NP-hard nature, finding an optimal solution for all scenarios within a reasonable time frame faces serious challenges. This paper proposes a solution that transforms the FJSSP into a Markov Decision Process (MDP) and employs deep reinforcement learning (DRL) techniques for resolution. First, we represent the state features of the scheduling environment using seven feature vectors and utilize a transformer encoder as a feature extraction module to effectively capture the relationships between state features and enhance representation capability. Second, based on the features of the jobs and machines, we design 16 composite dispatching rules from multiple dimensions, including the job completion rate, processing time, waiting time, and manufacturing resource utilization, to achieve flexible and efficient scheduling decisions. Furthermore, we project an intuitive and dense reward function with the objective of minimizing the total idle time of machines. Finally, to verify the performance and feasibility of the algorithm, we evaluate the proposed policy model on the Brandimarte, Hurink, and Dauzere datasets. Our experimental results demonstrate that the proposed framework consistently outperforms traditional dispatching rules, surpasses metaheuristic methods on larger-scale instances, and exceeds the performance of existing DRL-based scheduling methods across most datasets.

DLRover-RM: Resource Optimization for Deep Recommendation Models Training in the Cloud

Deep learning recommendation models (DLRM) rely on large embedding tables to manage categorical sparse features. Expanding such embedding tables can significantly enhance model performance, but at the cost of increased GPU/CPU/memory usage. Meanwhile, tech companies have built extensive cloud-based services to accelerate training DLRM models at scale. In this paper, we conduct a deep investigation of the DLRM training platforms at AntGroup and reveal two critical challenges: low resource utilization due to suboptimal configurations by users and the tendency to encounter abnormalities due to an unstable cloud environment. To overcome them, we introduce DLRover, an elastic training framework for DLRMs designed to increase resource utilization and handle the instability of a cloud environment. DLRover develops a resource-performance model by considering the unique characteristics of DLRMs and a three-stage heuristic strategy to automatically allocate and dynamically adjust resources for DLRM training jobs for higher resource utilization. Further, DLRover develops multiple mechanisms to ensure efficient and reliable execution of DLRM training jobs. Our extensive evaluation shows that DLRover reduces job completion times by 31%, increases the job completion rate by 6%, enhances CPU usage by 15%, and improves memory utilization by 20%, compared to state-of-the-art resource scheduling frameworks. DLRover has been widely deployed at AntGroup and processes thousands of DLRM training jobs on a daily basis. DLRover is open-sourced and has been adopted by 10+ companies.

Connection-Aware Heuristics for Scheduling and Distributing Jobs under Dynamic Dew Computing Environments

Due to the widespread use of mobile and IoT devices, coupled with their continually expanding processing capabilities, dew computing environments have become a significant focus for researchers. These environments enable resource-constrained devices to contribute computing power to a local network. One major challenge within these environments revolves around task scheduling, specifically determining the optimal distribution of jobs across the available devices in the network. This challenge becomes particularly pronounced in dynamic environments where network conditions constantly change. This work proposes integrating the “reliability” concept into cutting-edge human-design job distribution heuristics named ReleSEAS and RelBPA as a means of adapting to dynamic and ever-changing network conditions caused by nodes’ mobility. Additionally, we introduce a reinforcement learning (RL) approach, embedding both the notion of reliability and real-time network status into the RL agent. Our research rigorously contrasts our proposed algorithms’ throughput and job completion rates with their predecessors. Simulated results reveal a marked improvement in overall throughput, with our algorithms potentially boosting the environment’s performance. They also show a significant enhancement in job completion within dynamic environments compared to baseline findings. Moreover, when RL is applied, it surpasses the job completion rate of human-designed heuristics. Our study emphasizes the advantages of embedding inherent network characteristics into job distribution algorithms for dew computing. Such incorporation gives them a profound understanding of the network’s diverse resources. Consequently, this insight enables the algorithms to manage resources more adeptly and effectively.

The effects of big data analytics and workplace pressures on productivity

This study set out to find out on how big data analytics, workload, work wages, and organizational structure pressures affect employee job performance and overall business performance in the Jordan banking sector in the context of that country's banking sector. 292 samples from the research study were collected, analyzed, and debated to test the study's hypotheses. Employee and company success were found to be affected by workload, compensation, and organizational structure. Workload was reflected by job completion rate, average daily issues, and work completion time, while big data was reflected by external and internal, unique application, indexing, and source correctness. Work compensation, including employee compensation, employee experience and skill, and employee incentives and rewards, is the third primary factor. Communities of practice, group work, and physical infrastructure make up the fourth and last important factor, the organization's structure. The study finds that when dealing with large amounts of data, reducing the workload can improve employee performance, increasing employee motivation through higher wages and bonuses can improve performance, and having an organizational structure that promotes teamwork and work teams can improve employees' abilities to solve problems and improve work performance, which in turn affects overall productivity. This study contributes fresh information to an emerging field that requires more investigation to fully understand the interplay between big data, workload, work wages, and organizational structure pressures. The topic of this study, the Jordan banking sector, is both innovative and highly relevant because it may help financial institutions enhance their operations and provide superior customer service.

HunterPlus: AI based energy-efficient task scheduling for cloud–fog computing environments

Cloud computing is a mainstay of modern technology, offering cost-effective and scalable solutions to a variety of different problems. The massive shift of organization resource needs from local systems to cloud-based systems has greatly increased the costs incurred by cloud providers in expanding, maintaining, and supplying server, storage, network, and processing hardware. Due to the large scale at which cloud providers operate, even small performance degradation issues can cause energy or resource usage costs to rise dramatically. One way in which cloud providers may improve cost reduction is by reducing energy consumption. The use of intelligent task-scheduling algorithms to allocate user-deployed jobs to servers can reduce the amount of energy consumed. Conventional task scheduling algorithms involve both heuristic and metaheuristic methods. Recently, the application of Artificial Intelligence (AI) to optimize task scheduling has seen significant progress, including the Gated Graph Convolution Network (GGCN). This paper proposes a new approach called HunterPlus which examine the effect of extending the GGCN’s Gated Recurrent Unit to a Bidirectional Gated Recurrent Unit. The paper also studies the utilization of Convolutional Neural Networks (CNNs) in optimizing cloud–fog task scheduling. Experimental results show that the CNN scheduler outperforms the GGCN-based models in both energy consumption per task and job completion rate metrics by at least 17 and 10.4 percent, respectively.

Qore-DL: A QoS-aware joint optimization framework for distributed deep learning training

Resource management of deep learning training (DLT) jobs is critical for cluster resource efficiency and client QoS assurance. Most existing scheduling frameworks require clients to specify job resource configuration, which can lead to over-provision or under-provision issues. Additionally, the performance of some static scheduling frameworks degrades in highly dynamic clusters. In this paper, we propose a QoS-aware joint resource optimization framework called Qore-DL for distributed DLT jobs. We divide the lifecycle of a DLT job into submission, queuing and running stages. Qore-DL automatically configures reasonable resources for submitted jobs and greedily assigns scheduled jobs to hosts. For running jobs, Qore-DL employs a heuristic scheme to adjust their resources. Qore-DL jointly considers the optimization of QoS satisfaction and resource efficiency at the three stages of DLT jobs. We implemented the prototype of Qore-DL in TensorFlow based on Kubernetes and conducted extensive experiments in CPU and GPU clusters to evaluate its performance. The experiment results show that, compared with its counterparts, Qore-DL can improve the job completion rate by up to 42.4% and the cluster resource efficiency by up to 21.8%.

Dynamic power management for value-oriented schedulers in power-constrained HPC system

High performance computing (HPC) systems are confronting the challenge of improving their productivity under a system-wide power constraint in the exascale era. To measure the productivity of an HPC job, researchers have proposed to assign a monotonically decreasing time-dependent value function, called job-value, to that job. These job-value functions are used by the value-based scheduling algorithms to maximize the system productivity where system productivity is the accumulation of job-value for the completed jobs. In this study, we first show that the relative performance of the competing state-of-the-art static power allocation strategies interchange based on the level of the power constraint when applied to the value-based algorithms. We then investigate the limitations of these static strategies by relating the job completion rate to the resource utilization, and expose that there is non-negligible amount of unused resources for the scheduler to utilize. Even though the system is oversubscribed, these unused resources are insufficient to schedule new high-value jobs. Based on this observation, we propose a novel dynamic power management strategy for the value-based algorithms. Our dynamic allocation policy maximizes the system productivity, resource utilization, and job completion rate by utilizing application power-performance models to reallocate power from running jobs to newly arrived jobs. We simulate a large-scale system that uses job arrival traces from a real HPC system. We demonstrate that the dynamic-variant of each value-based algorithm earns up to 16% higher productivity and completes 13% more jobs compared to its static variants when power becomes a highly constrained resource in the system.

Ant Colonized and Taguchi Parallel Scheduliing with Sequence Independent Setup Time

The Job Shop Parallel Machine Scheduling (JSPMS) is a hybrid production system, and hence has received significant attention in the past few years. The JSPMS problem is a rationalization of the traditional job shop scheduling problem in computer science and operation research that permits to process operations on single machine out of a set of possible parallel machines. To maximize the job completion rate and minimize job completion time, a hybrid production system is necessary. With this objective, a novel meta-heuristic method is designed. This paper develops a scheduling method called, Ant Colonized and Taguchi Parallel Operation Scheduling (ACTPOS), for JSPMS, aimed to minimize job completion time and maximize job completion rate. The design of AC-TPOS method involves two different models, namely, Ant Colonized Parallel Machine Selection (ACPMS) model and Taguchi Parallel Operation Scheduling (TPOS) model. In ACPMS model, optimal selection of machine is done via operation being processed by parallel machines using local pheromone updating rule concentrating on the makespan time. In addition, the processing time and sequence-independent setup time are considered. Next, in TPOS model, optimal scheduling of operation is performed using Taguchi method concentrating on the makespan rate. Finally, the test results first show that our algorithm outperforms existing methods in terms of job completion rate, job completion time and computational complexity involved in scheduling operations.

Computational Load Balancing on the Edge in Absence of Cloud and Fog

Mobile Cloud Computing or Fog computing refers to offloading computationally intensive algorithms from a mobile device to the cloud or an intermediate cloud in order to save resources e.g., time and energy in the mobile device. This paper proposes new solutions for situations when the cloud or fog is not available. First, the sensor network is modelled using a network of queues, then a linear programming technique is used to make scheduling decisions. Various centralized and distributed algorithms are then proposed, which improves overall system performance. Extensive simulations show slightly higher energy usage in comparison to the baseline non-offloading case, however, the job completion rate is significantly improved, the efficiency score metric shows the extra energy usage is justified. The algorithms have been simulated in various environments including high and low bandwidth, partial connectivity, and different rate of information exchanges to study the pros and cons of the proposed algorithms.

Shared-object system equilibria: Delay and throughput analysis

We consider shared-object systems that require their threads to fulfill the system jobs by first acquiring sequentially the objects needed for the jobs and then holding on to them until the job completion. Such systems are in the core of a variety of shared-resource allocation and synchronization systems. This work opens a new perspective to study the expected job delay and throughput analytically, given the possible set of jobs that may join the system dynamically. We identify the system dependencies that cause contention among the threads as they try to acquire the job objects. We use these observations to define the shared-object system equilibria. We note that the system is in equilibrium whenever the rate in which jobs arrive at the system matches the job completion rate. These equilibria consider not only the job delay but also the job throughput, as well as the time in which each thread blocks other threads in order to complete its job. We then further study in detail the thread work cycles and, by using a graph representation of the problem, we are able to propose procedures for finding and estimating equilibria, i.e., discovering the job delay and throughput, as well as the blocking time. To the best of our knowledge, this is a new perspective, that can provide better analytical tools for the problem. That is, our methods can be used to estimate performance measures similar to ones that can be acquired through experimentation on working systems and simulations, e.g., as job delay and throughput in (distributed) shared-object systems.

Evaluation and optimization of the mixed redundancy strategy in cloud-based systems

Mixed redundancy strategies are generally used in cloud-based systems, with different node switch mechanisms from traditional fault-tolerant strategies. Existing studies often concentrate on optimizing a single strategy in cloud computing environment and ignore the impact of mixed redundancy strategies. Therefore, a model is proposed to evaluate and optimize the reliability and performance of cloud-based degraded systems subject to a mixed active and cold standby redundancy strategy. In this strategy, node switching is triggered by a continual monitoring and detection mechanism when active nodes fail. To evaluate the transient availability and the expected job completion rate of systems with such kind of strategy, a continuous-time Markov chain model is built on the state transition process and a numerical method is used to solve the model. To choose the optimal redundancy for the mixed strategy under system constraints, a greedy search algorithm is proposed after sensitivity analysis. Illustrative examples were presented to explain the process of calculating the transient probability of each system state and in turn, the availability and performance of the whole system. It was shown that the near-optimal redundancy solution could be obtained using the optimization method. The comparison with optimization of the traditional mixed redundancy strategy proved that the system behavior was different using different kinds of mixed strategies and less redundancy was assigned for the new type of mixed strategy under the same system constraint.

A cybernetics Social Cloud

The SocialMedia API has been developed to demonstrate proofs-of-concept.Cybernetics-1 and cybernetics-2 functions are developed for cybernetics Social Cloud.Large scale simulations have been tested for capacity tests. No costs are involved.All the steps involved are fully justified for research contributions. This paper proposes a Social Cloud, which presents the system design, development and analysis. The technology is based on the BOINC open source software, our hybrid Cloud, Facebook Graph API and our development in a new Facebook API, SocialMedia. The creation of SocialMedia API with its four functions can ensure a smooth delivery of Big Data processing in the Social Cloud, with four selected examples provided. The proposed solution is focused on processing the contacts who click like or comment on the author's posts. Outputs result in visualization with their core syntax being demonstrated. Four functions in the SocialMedia API have evaluation test and each client-server API processing can be completed efficiently and effectively within 1.36źs. We demonstrate large scale simulations involved with 50,000 simulations and all the execution time can be completed within 70,000źs. Cybernetics functions are created to ensure that 100% job completion rate for Big Data processing. Results support our case for Big Data processing on Social Cloud with no costs involved. All the steps involved have closely followed system design, implementation, experiments and validation for Cybernetics to ensure a high quality of outputs and services at all times. This offers a unique contribution for Cybernetics to meet Big Data research challenges.

A Hybrid Approach for Scheduling Virtual Machines in Private Clouds

Quality of Service (QoS) support in private clouds is a challenging process because of the limitations of available resources and the high rate of received jobs, which leads to an NP hard scheduling problem. In private clouds, resource owners are usually interested in maximizing their resource utilization and completion rates while minimizing the turnaround time of their jobs, which complicates the scheduling problem even more. Haizea is an eminent cloud scheduler that offers high performance in terms of job turnaround time and completion rate. However, Haizea, and cloud schedulers in general, suffer from low resource utlization. Additionally, cloud schedulers usually consider only end users’ demands, while providers’ demands are entirely neglected. This is because an infinite pool of resources is assumed, which is difficult to achieve and simply not true in private clouds. Conversely, Condor, the eminent High Throughput Computing (HTP) scheuler, is known for addressing these shortcomings by formulating owner's and user's requirements as a logical expression evaluated based on the context which result is high resource utilization. Unfortunatly, this comes with the price of long execution time. As each of Haizea and Condor has its own advantages and limitations, in this paper, we propose a hybrid Haizea and Condor approach (HHCS) which utilizes techniques from both schedulers in a way that maximizes their advantages and overcomes their limitations. The proposed approach has been tested thoroughly in a simulated private cloud environment under various numbers of nodes and jobs. Experimental results illustrated an enhanced performance in terms of resources utilization without compromising the job turnaround time or the job completion rate.

Scalability and interoperability within glideinWMS

Physicists have access to thousands of CPUs in grid federations such as OSG and EGEE. With the start-up of the LHC, it is essential for individuals or groups of users to wrap together available resources from multiple sites across multiple grids under a higher user-controlled layer in order to provide a homogeneous pool of available resources. One such system is glideinWMS, which is based on the Condor batch system. A general discussion of glideinWMS can be found elsewhere. Here, we focus on recent advances in extending its reach: scalability and integration of heterogeneous compute elements. We demonstrate that the new developments exceed the design goal of over 10,000 simultaneous running jobs under a single Condor schedd, using strong security protocols across global networks, and sustaining a steady-state job completion rate of a few Hz. We also show interoperability across heterogeneous computing elements achieved using client-side methods. We discuss this technique and the challenges in direct access to NorduGrid and CREAM compute elements, in addition to Globus based systems.

Failure-aware resource management for high-availability computing clusters with distributed virtual machines

In large-scale networked computing systems, component failures become norms instead of exceptions. Failure-aware resource management is crucial for enhancing system availability and achieving high performance. In this paper, we study how to efficiently utilize system resources for high-availability computing with the support of virtual machine (VM) technology. We design a reconfigurable distributed virtual machine (RDVM) infrastructure for networked computing systems. We propose failure-aware node selection strategies for the construction and reconfiguration of RDVMs. We leverage the proactive failure management techniques in calculating nodes’ reliability states. We consider both the performance and reliability status of compute nodes in making selection decisions. We define a capacity–reliability metric to combine the effects of both factors in node selection, and propose Best-fit algorithms with optimistic and pessimistic selection strategies to find the best qualified nodes on which to instantiate VMs to run user jobs. We have conducted experiments using failure traces from production systems and the NAS Parallel Benchmark programs on a real-world cluster system. The results show the enhancement of system productivity by using the proposed strategies with practically achievable accuracy of failure prediction. With the Best-fit strategies, the job completion rate is increased by 17.6% compared with that achieved in the current LANL HPC cluster. The task completion rate reaches 91.7% with 83.6% utilization of relatively unreliable nodes.

Multiple period repair kit problem with a job completion criterion: A case study

The repair kit problem is concerned with finding an optimal kit of parts and tools to carry for on-site repairs. This choice involves a trade-off between the annual holding cost and a penalty for failing to complete repairs during the first visit. The repair kit problem with job completion criterion tries to find a solution to this trade-off by achieving a predetermined job completion rate (or service level) with minimum holding costs. This formulation of the problem avoids the often difficult definition and calculation of penalty costs. The repair kit problem has been studied in literature for a single period, when restocking is possible in between jobs. In the multiple period repair kit problem studied here, restocking of the spare parts is only possible every 12 periods. Other assumptions used in this case study are that failures of the various parts are stochastically independent and that at most one part of each type is used on a given job. This paper describes a case study where the multiple period repair kit problem with a job completion criterion was solved using a knapsack heuristic. First the model for this problem is developed, then numerical results are shown and conclusions drawn.

Optimal inventories for repairable redundant systems with aging components

AbstractComplex systems that are required to perform very reliably are often designed to be “fault‐tolerant,” so that they can function even though some component parts have failed. Often fault‐tolerance is achieved through redundancy, involving the use of extra components. One prevalent redundant component configuration is the m‐out‐of‐n system, where at least m of n identical and independent components must function for the system to function adequately.Often machines containing m‐out‐of‐n systems are scheduled for periodic overhauls, during which all failed components are replaced, in order to renew the machine's reliability. Periodic overhauls are appropriate when repair of component failures as they occur is impossible or very costly. This will often be the case for machines which are sent on “missions” during which they are unavailable for repair. Examples of such machines include computerized control systems on space vehicles, military and commercial aircraft, and submarines.An interesting inventory problem arises when periodic overhauls are scheduled. How many spare parts should be stocked at the maintenance center in order to meet demands? Complex electronic equipment is rarely scrapped when it fails. Instead, it is sent to a repair shop, from which it eventually returns to the maintenance center to be used as a spare. A Markov model of spares availability at such a maintenance center is developed in this article. Steady‐state probabilities are used to determine the initial spares inventory that minimizes total shortage cost and inventory holding cost. The optimal initial spares inventory will depend upon many factors, including the values of m and n, component failure rate, repair rate, time between overhauls, and the shortage and holding costs.In a recent paper, Lawrence and Schaefer [4] determined the optimal maintenance center inventories for fault‐tolerant repairable systems. They found optimal maintenance center inventories for machines containing several sets of redundant systems under a budget constraint on total inventory investment. This article extends that work in several important ways. First, we relax the assumption that the parts have constant failure rates. In this model, component failure rates increase as the parts age. Second, we determine the optimal preventive maintenance policy, calculating the optimal age at which a part should be replaced even if it has not failed because the probability of subsequent failure has become unacceptably high. Third, we relax the earlier assumption that component repair times are independent, identically distributed random variables. In this article we allow congestion to develop at the repair shop, making repair times longer when there are many items requiring repair. Fourth, we introduce a more efficient solution method, marginal analysis, as an alternative to dynamic programming, which was used in the earlier paper. Fifth, we modify the model in order to deal with an alternative objective of maximizing the job‐completion rate.In this article, the notation and assumptions of the earlier model are reviewed. The requisite changes in the model development and solution in order to extend the model are described. Several illustrative examples are included.

Note—On “Optimizing Field Repair Kits Based on Job Completion Rate”

In a recent paper, Mamer and Smith (Mamer, J. W., S. A. Smith. 1982. Optimizing field repair kits based on job completion rate. Management Sci. 28 (11) 1328–1333.) presented a formulation for determining the optimal kit of parts and tools for on-site equipment repairs under a job completion criterion. The optimal kit minimizes the sum of expected broken job costs plus the annual inventory holding cost for those parts stocked. A broken job occurs whenever a repair cannot be completed because one or more of its required parts are not included in the field repair kit. Assuming equal penalties for all broken jobs, an algorithm developed by Eisner and Severance (Eisner, M. J., D. G. Severance. 1976. Mathematical techniques for efficient record segmentation in large shared databases. J. Assoc. Comput. Mach. 23 (4) 619–635.) in a different context can be used to determine a set of optimal policies and a range on the penalty cost for which each policy is optimal. The ranges cover all penalty costs from 0 to ∞. This is significant for the repair kit problem since the actual penalty cost of broken jobs may be difficult to determine exactly.

A Multi-Item Maintenance Center Inventory Model for Low-Demand Reparable Items

In many military and commercial contexts, complex equipment undergoes scheduled maintenance overhauls at regular intervals during which all failed components are replaced. Failure to have replacements on hand for all failed parts requires emergency measures at premium cost. When reparable parts are highly reliable and expensive, both holding and shortage costs are high. This model determines the reparable parts inventory for a maintenance center under three alternative criteria: (1) maximizing job-completion rate subject to constraint on total holding costs, (2) minimizing total holding costs plus expected job-noncompletion costs, and (3) minimizing total holding costs subject to a required minimum job-completion rate. Exact solutions may be obtained using dynamic programming. Approximate solutions, found easily by marginal analysis, have readily computed bounds on possible error. The solution methods for the three formulations are illustrated in a simple example.

Optimizing Field Repair Kits Based on Job Completion Rate

A formulation is presented for obtaining the optimal kit of parts and tools for on-site equipment repairs, based on the fraction of jobs which can be completed from the contents of the kit. This formulation generalizes previous approaches in that part demands need not be independent and various numbers of parts as well as multiple part types per job may be required. It is shown that the formulation is a special case of Balinski's “selection problem” and thus can be solved by maximal flow/minimal cut network algorithms.