Job Processing Research Articles

An improved multi-objective evolutionary algorithm based on decomposition for solving re-entrant hybrid flow shop scheduling problem with batch processing machines

Cold-drawn seamless steel pipes play a pivotal role in special industries. However, scheduling plan, as the core of the production, is manually made, which leads to the issue of low production efficiency. The change in processing time or job families when the jobs reenter and the appropriate buffer size of batch processing machines are not considered in the current research. The energy consumption is not considered as an important objective as well. These gaps motivate this study. The production process of cold-drawn seamless steel pipes can be defined as a re-entrant hybrid flow shop scheduling problem with batch processing machines. The model is formulated to minimize the makespan and the energy consumption of the batch processing machines. An improved multi-objective evolutionary algorithm based on decomposition is proposed to address this problem. To make an energy-efficient production scheduling, the buffer size of batch processing machines is discussed and the batching method is developed. Besides, a greedy selection strategy is proposed to choose a machine for cold-drawn operation considering the influence of the long processing time. The results of experiments show that the appropriate buffer size is determined to be the same as the capacity of batch processing machines. The proposed greedy selection strategy can select suitable machines for jobs. The results prove the effectiveness of the batching constraint and the balanced decoding algorithm, and the proposed algorithm can effectively solve the re-entrant hybrid flow shop scheduling problem with batch processing machines.

A reinforcement learning based RMOEA/D for bi-objective fuzzy flexible job shop scheduling

The flexible job shop scheduling problem (FJSP) is significant for realistic manufacturing. However, the job processing time usually is uncertain and changeable during manufacturing. This paper presents a multi-objective FJSP with fuzzy processing time (MOFFJSP) for optimizing the makespan and total machine workload as objectives. To solve the MOFFJSP, a MOEA/D based on reinforcement learning named RMOEA/D is proposed. RMOEA/D can be featured as: (i) an initial strategy with three rules is used to get a high-quality initial population; (ii) a parameter adaption strategy based on Q-learning is proposed to guide the population choose the best parameter to increase diversity; (iii) a variable neighborhood search based on reinforcement learning is designed to lead the solution to choose the right local search method; and (iv) an elite archive is used to improve the usage rate of the abandoned historical solution. RMOEA/D is compared with five well-known realted methods, i.e., MOEA/D, NSGA-II, MOEA/D-M2M, NSGA-III and IAIS on three benchmark suites. The results show that RMOEA/D outperforms these five state-of-art algorithms.

Occupational Benzene Exposure in the Norwegian Offshore Petroleum Industry, 2002-2018.

PurposeWorkers on offshore petroleum installations are at risk of being exposed to benzene which is carcinogenic to humans. The present study aimed to assess the time trend of full-shift benzene exposure from 2002 to 2018 in order to characterize benzene exposure among laboratory technicians, mechanics, process operators, and industrial cleaners, and to examine the possible determinants of benzene exposure.MethodsA total of 924 measurements of benzene exposure from the Norwegian petroleum offshore industry were included. The median sampling duration was 680 min, ranging from 60 to 940 min. The overall geometric mean (GM) and 95% confidence interval, time trends, and determinants of exposure were estimated using multilevel mixed-effects tobit regression analyses. Time trends were estimated for sampling duration below and above 8 h, both overall and for job groups. The variability of exposure between installation and workers was investigated in a subset of data containing worker identification.ResultsThe overall GM of benzene exposure was 0.004 ppm. When adjusting for job group, design of process area, season, wind speed, and sampling duration, industrial cleaners had the highest exposure (GM = 0.012). Laboratory technicians, mechanics, and process operators had a GM exposure of 0.004, 0.003, and 0.004 ppm, respectively. Overall, the measured benzene exposure increased by 7.6% per year from 2002 to 2018. Mechanics had an annual increase of 8.6% and laboratory technicians had an annual decrease of 12.6% when including all measurements. When including only measurements above 8 h, mechanics had an increase of 16.8%. No statistically significant time trend was found for process operators. Open process area, high wind speed, and wintertime were associated with reduced exposure level.ConclusionsAn overall increase in measured exposure was observed from 2002 to 2018. The increase may reflect changes in measurement strategy from mainly measuring on random days to days with expected exposure. However, the time trend varied between job groups and was different for sampling duration above or below 8 h. Industrial cleaners had the highest exposure of the four job groups while no differences in exposure were observed between laboratory technicians, mechanics, and process operators. The design of the process area, job group, wind speed, and season were all significant determinants of benzene exposure.

Single machine group scheduling problem with makespan objective and a proportional linear shortening

This paper considers a group scheduling problem with shorten (i.e., a proportional linear shortening) job processing times and ready times on a single machine. If the jobs are in the same group, they have independent and constant ready times. The setup time of a group is independent constant setup time between each group. The goal is to determine the optimal group sequence and the job sequence within the groups such that the makespan (i.e., the maximum completion time) is minimized. For the general case of the problem, an initial heuristic algorithm (an upper bound) and some lower bounds are proposed, and then a branch-and-bound algorithm can be developed to solve the problem.

특성화고 교장의 직무수행에 대한 현상학적 연구

The purpose of this study was to analysis job performance of specialized vocational high school principals using phenomenological research method. The interview for the research was carried out with 12 specialized vocational high school principals from November 29 to December 30, 2021.
 According to the result of the research, 27 themes were extracted from interviews with 12 research participants. And 10 theme clusters were categorized from these.
 Based on the finding of the study, major conclusions of this study were as follow:First, specialized vocational high school principals have following attitude for job. ‘to make a commitment to revive vocational education’, ‘to listen intently, to empathize, to feel empathy’, ‘to lower authority and power’, ‘to consider others’, ‘to lead to cooperation’, ‘to suggest right direction and vision’, ‘to consider students first’, ‘to decide reasonably’.
 Second, they experience following difficulties when processing job performance. ‘to boost morale’, ‘to encourage teachers’, ‘to mediate generational conflict between teachers’, ‘to execute budget’, ‘to promote various business’.
 Third, in order for them to perform job, they need following supports. ‘to complement job training program’, ‘to expand autonomy of specialized vocational high school principal’, ‘to provide institutional support’.

Mengeksplorasi Perubahan Sifat Kerja Dan Kinerja ASN Pemerintah Daerah Jawa Timur Pasca Covid-19

Local governments should maximize employee to produce improved outcomes and efficiency. Performance is related to the result and how the job process is carried out since it is this process that determines a person's level of achievement at work. What needs to be understood is how the Civil Servants' efficiency can be optimized, and leaders must pay attention to enhancing their performance. This study aims to examine the effect the of change nature of work on the performance of Indonesian local government civil servants in the post-Covid-19 period. The result shows that to build good collaboration, it is necessary to understand between employees so that no miscommunication and leaders can manage subordinates well with attitude deep concern for their employee. It also requires high responsiveness, responsibility, and transparency in decision making. Besides that, it is necessary to understand the importance of cooperation to providing socialization to all employees.

Single-machine scheduling with total late work and job rejection

We study single-machine scheduling problems with rejection, where we can refuse to process jobs in the shop at a certain cost. Solutions to the problems investigated in this study consist of (i) partitioning the set of jobs into a set of accepted and a set of rejected jobs, and (ii) scheduling the set of accepted jobs on the single machine. The quality of a solution is measured by two criteria. The first is the total late work of the set of accepted jobs, and the second is the total rejection cost. We study two problems. The goal in the first is to find a solution minimizing the sum of the total late work and the total rejection cost, while the goal in the second is to find a solution minimizing the total rejection cost, given a bound on the total late work. As both problems are NP -hard, we focus on the design of pseudo-polynomial time algorithms and approximation schemes.

Statistical analysis of multi job processing in Hadoop environment using schedulers

Hadoop schedulers facilitate effective sharing of system resources among multiple tasks for the purpose of effective completion of jobs. Upon submission of a job, the same will be transferred onto a queue by Hadoop scheduler. Once the transfer is complete, aforesaid job will be split up into multiple tasks and is further distributed among various computing nodes. Herein, performance is directly related to time taken up by aforesaid tasks to complete the assigned job. Optimum distribution of system resources among such tasks play a vital role in reduction of job completion time, and thus can ensure improved performance in relation to the same. Hadoop, by virtue uses FIFO scheduler. In the proposed experiment, FIFO scheduler’s job execution time is discussed and compared with the job execution time of Fair scheduler and Capacity scheduler.

두 단계 흐름생산라인에서 가중완료시간을 최소화하기 위해 아웃소싱 전략을 활용하는 일정계획문제

This paper researches a two-machine flowshop scheduling with an outsourcing strategy. In the flowshop, the first processing of jobs is either managed in-house or outsourced to outside suppliers, the second processing is processed in-house only. When a job is outsourced, the company pays outsourcing cost for each job, and considers outsourcing lead-time implying that the second processing can start after the lead-time. The objective cost function of this paper is the sum of the total weighted completion time and the outsourcing costs. This paper derives a branch-and-bound algorithm and some heuristics for the problem. From the computational experiments, we can observe that the performance of the derived algorithms is effective and efficient.

A target-based distributionally robust model for the parallel machine scheduling problem

ABSTRACT We develop a distributionally robust optimisation (DRO) model based on a risk measure for the parallel machine scheduling problem (PMSP) with random job processing times. We propose an underperformance risk index (URI) to control the extent of the total weighted completion time (TWCT) that exceeds target level T. With partially characterised uncertainty set information, we transform the model with URI to its equivalent mixed-integer linear programming (MILP) counterparts. Due to the NP-hardness of PMSP with different job weights, we design a hybrid algorithm with a heuristic assignment and exact subproblem for large-scale problems. The proposed hybrid algorithm reduces the computation time significantly at the expense of solution quality. We also introduce a reformulation approach under the setting of equally weighted and identical machines. Numerical results show that our model performs better than the distributionally β-robust optimisation models. Our proposed URI accounts for both the frequency and magnitude of violation from the target. The uncertainty set we used preserves a linear structure under partially characterised distributional information. Our computational results and sensitivity analysis show the effectiveness and efficiency of our proposed DRO model under various settings, including different problem sizes, different processing time variations, and information misalignment.

The Case for Phase-Aware Scheduling of Parallelizable Jobs

Parallelizable workloads are ubiquitous and appear across a diverse array of modern computer systems. Data centers, supercomputers, machine learning clusters, distributed computing frameworks, and databases all process jobs designed to be parallelized across many servers or cores. Unlike the jobs in more classical models, such as the M/G/k queue, that each run on a single server, parallelizable jobs are capable of running on multiple servers simultaneously. When a job is parallelized across additional servers or cores, the job receives a speedup and can be completed more quickly.

Deep Reinforcement Learning for Distributed Flow Shop Scheduling with Flexible Maintenance

A common situation arising in flow shops is that the job processing order must be the same on each machine; this is referred to as a permutation flow shop scheduling problem (PFSSP). Although many algorithms have been designed to solve PFSSPs, machine availability is typically ignored. Healthy machine conditions are essential for the production process, which can ensure productivity and quality; thus, machine deteriorating effects and periodic preventive maintenance (PM) activities are considered in this paper. Moreover, distributed production networks, which can manufacture products quickly, are of increasing interest to factories. To this end, this paper investigates an integrated optimization of the distributed PFSSP with flexible PM. With the introduction of machine maintenance constraints in multi-factory production scheduling, the complexity and computation time of solving the problem increases substantially in large-scale arithmetic cases. In order to solve it, a deep Q network-based solution framework is designed with a diminishing greedy rate in this paper. The proposed solution framework is compared to the DQN with fixed greedy rate, in addition to two well-known metaheuristic algorithms, including the genetic algorithm and the iterated greedy algorithm. Numerical studies show that the application of the proposed approach in the studied production-maintenance joint scheduling problem exhibits strong solution performance and generalization abilities. Moreover, a suitable maintenance interval is also obtained, in addition to some managerial insights.

Realtime scheduling heuristics for just-in-time production in large-scale flexible job shops

This study aims to enable jobs to go smoothly between shops on a production line by completing jobs in the upstream shop just in time (JIT) for the downstream shop. We propose solutions to a factory that is seeking ways for an upstream shop to complete every job at the precise time such that the downstream shop can process the job. We model the upstream shop as a flexible job shop and propose four methods that form a realtime scheduling and control system for JIT production. We first propose a method to set for each job a due date by which the job should be completed in the upstream shop. The due dates are set in such a manner that jobs would be completed JIT for the downstream shop, if they are completed JIT for their due dates. We then propose a method to estimate the minimum number of workers needed in the upstream shop for completing the jobs by their due dates. We further propose two methods that work dynamically to complete each job neither too early nor too late for its due date. One is a dispatching rule that dynamically sequences jobs in process according to urgency degree. The other is a job-selecting heuristic that dynamically assigns workers to jobs such that jobs not nearing completion will be given priority in processing. Simulations by using data from the factory show that the methods can achieve in real time (i.e. within 0.00 seconds) JIT production for a flexible job shop problem involving hundreds of operations. More extensive simulations by using a large number of randomly generated problem instances show that solutions obtained in real time by the proposed methods greatly outperform those obtained in much longer time by metaheuristics designed for solving similar problems, and that each proposed method outperforms its rivals in the literature. The findings imply that integrating fast and high-performing heuristics and rules can be a solution to solve large-scale scheduling problems in real time.

An effective new heuristic algorithm for solving permutation flow shop scheduling problem

The deterministic permutation flow shop scheduling problem with makespan criterion is not solvable in polynomial time‎. ‎Therefore‎, ‎researchers have thought about heuristic algorithms‎. ‎There are many heuristic algorithms for solving it that is a very important combinatorial optimization problem‎. ‎In this paper‎, ‎a new algorithm is proposed for solving the mentioned problem‎. ‎The presented algorithm chooses the weighted path that starts from the up-left corner and reaches the down-right in the matrix of jobs processing times and calculates the biggest sum of the times in the footprints of this path‎. ‎The row with the biggest sum permutes among all the rows of the matrix for locating the minimum of makespan‎. ‎This method was run on Taillard’s standard benchmark and the solutions were compared with the optimum or the best ones as well as 14 famous heuristics‎. ‎The validity and effectiveness of the algorithm are shown with tables and statistical evaluation‎.

A Comprehensive Empirical Study of Query Performance Across GPU DBMSes

In recent years, GPU database management systems (DBMSes) have rapidly become popular largely due to their remarkable acceleration capability obtained through extreme parallelism in query evaluations. However, there has been relatively little study on the characteristics of these GPU DBMSes for a better understanding of their query performance in various contexts. Also, little has been known about what the potential factors could be that affect the query processing jobs within the GPU DBMSes. To fill this gap, we have conducted a study to identify such factors and to propose a structural causal model, including key factors and their relationships, to explicate the variances of the query execution times on the GPU DBMSes. We have also established a set of hypotheses drawn from the model that explained the performance characteristics. To test the model, we have designed and run comprehensive experiments and conducted in-depth statistical analyses on the obtained empirical data. As a result, our model achieves about 77% amount of variance explained on the query time and indicates that reducing kernel time and data transfer time are the key factors to improve the query time. Also, our results show that the studied systems should resolve several concerns such as bounded processing within GPU memory, lack of rich query evaluation operators, limited scalability, and GPU under-utilization.

A Review Paper on Work-Balance Life in Large Organization

There is a growing readiness in today's workplace that workers are not sacrificing their lives just because they are working. Work and health are always two of the most important aspects of a single married life. However, the process of processing job and non-employment applications is one of the modern ways of paying attention to both individuals and organizations. With the development of the diversity of family structures mentioned in today's workforce, especially with the evolving level of dual work families, the importance of tackling the working life balance has grown significantly in recent years. Managers understand that the immediate living environment of the employee and the home and family affects the quality of the work and that there are strong business motives for job advertising and non-work communication. In this project, we strive that helping employees achieve a working life balance should become a fundamental part of HR policy and practice in the event that it is to get the best out of the organisation's relatives without throwing them dissatisfied, old and dissatisfied, unfulfilled. Keywords: Employees work balance, work life balance, work life stability

The Dynamic Flow Shop Scheduling Problem with JIT Philosophy and Common Due Date

Most studies on common due date problems discuss the topic in the context of solving the minimum penalty value for the single-machine or parallel machine scheduling environment. This study extends the problem of the common due date to the dynamic flow shop environment and proposes the enhanced heuristic algorithm to solve the minimum penalty value. The enhanced heuristic algorithm is characterized by designating mutant genomes of the child as the genomes located at the central location before mutation. The advantage is to integrate the successful experiences of the conventional common due date algorithm to improve the efficiency of the proposed heuristic algorithm. The performance of both algorithms is analyzed in terms of uniformly distributed job numbers, processing time, early penalty, and late penalty. The resulting outcomes indicate that the enhanced heuristic algorithm outperforms the conventional CDDA and EDD proposed by previous studies in the average penalties.

A branch‐and‐price approach to the multitasking scheduling with batch control on parallel machines

AbstractThis paper addresses multitasking scheduling with batch control on parallel machines. The specific scenario can be found in the maintenance activities of skilled workers. The objective is to minimize the total weighted completion time. In the problem, each machine needs to be allocated at least to one batch of jobs. On each machine, the processing of a primary job may be interrupted by the waiting jobs of the same machine, which can be described as the multitasking scheduling. The existing algorithms and mathematical formulations of such a scheduling problem are not suitable for the problem. An exact branch‐and‐price algorithm with heuristic pricing strategy is devised. To fast find the minimized reduced cost of the pricing problem, an artificial bee colony algorithm is developed. Specifically, it is aimed at obtaining a partitioning schedule with a minimized reduced cost, which involves the allocation of batches and the multitasking scheduling of jobs. A weak enumeration operator is proposed to cope with the batch allocation. The computational results demonstrate that the branch‐and‐price algorithm with heuristic pricing strategy outperforms general‐purpose pure CPLEX solver by orders of magnitude.

Cost Efficient GPU Cluster Management for Training and Inference of Deep Learning

Expanding the scale of GPU-based deep learning (DL) clusters would bring not only accelerated AI services but also significant energy consumption costs. In this paper, we propose a cost efficient deep learning job allocation (CE-DLA) approach minimizing the energy consumption cost for the DL cluster operation while guaranteeing the performance requirements of user requests. To do this, we first categorize the DL jobs into two classes: training jobs and inference jobs. Through the architecture-agnostic modeling, our CE-DLA approach is able to conduct the delicate mapping of heterogeneous DL jobs to GPU computing nodes. Second, we design the electricity price-aware DL job allocation so as to minimize the energy consumption cost of the cluster. We show that our approach efficiently avoids the peak-rate time slots of the GPU computing nodes by using the sophisticated mixed-integer nonlinear problem (MINLP) formulation. We additionally integrate the dynamic right-sizing (DRS) method with our CE-DLA approach, so as to minimize the energy consumption of idle nodes having no running job. In order to investigate the realistic behavior of our approach, we measure the actual output from the NVIDIA-based GPU devices with well-known deep neural network (DNN) models. Given the real trace data of the electricity price, we show that the CE-DLA approach outperforms the competitors in views of both the energy consumption cost and the performance for DL job processing.

Some Scheduling Problems With Job Rejection And A Learning Effect

AbstractIn this paper, we examined single and parallel machine scheduling problems with a learning effect and job rejection simultaneously. In real life, job processing times decrease when there is a learning effect. In some cases, producers cannot process all the jobs and pay the penalty cost for these jobs that they do not process. In our study, learning effect and job rejection are considered at the same time. We examined four different objective functions. Our objectives for single-machine scheduling problems are makespan and rejection cost minimization, total completion time and rejection cost minimization and total absolute deviation of completion times (TADC) and rejection cost minimization. Our objective for parallel machines is makespan and rejection cost minimization. The problems are solved by mathematical models, and four different algorithms are proposed for the problems. From these algorithms, the same results are obtained with single-machine makespan and rejection cost minimization, parallel machine makespan and rejection cost minimization and total completion time and rejection cost minimization. The accuracy for these models is obtained as 100%. The proposed algorithm for TADC and rejection cost minimization yielded close-to-optimal results. Mathematical model and algorithm results for 10 jobs, 20 jobs and 30 jobs are compared and the results are presented. The obtained solutions are obtained in polynomial time.