Workload Balancing Research Articles

Intelligent multi-agent reinforcement learning model for resources allocation in cloud computing

Now more than ever, optimizing resource allocation in cloud computing is becoming more critical due to the growth of cloud computing consumers and meeting the computing demands of modern technology. Cloud infrastructures typically consist of heterogeneous servers, hosting multiple virtual machines with potentially different specifications, and volatile resource usage. This makes the resource allocation face many issues such as energy conservation, fault tolerance, workload balancing, etc. Finding a comprehensive solution that considers all these issues is one of the essential concerns of cloud service providers. This paper presents a new resource allocation model based on an intelligent multi-agent system and reinforcement learning method (IMARM). It combines the multi-agent characteristics and the Q-learning process to improve the performance of cloud resource allocation. IMARM uses the properties of multi-agent systems to dynamically allocate and release resources, thus responding well to changing consumer demands. Meanwhile, the reinforcement learning policy makes virtual machines move to the best state according to the current state environment. Also, we study the impact of IMARM on execution time. The experimental results showed that our proposed solution performs better than other comparable algorithms regarding energy consumption and fault tolerance, with reasonable load balancing and respectful execution time.

Understanding postal delivery areas in the Republic of Korea using multiple unsupervised learning approaches

Changes in household composition and the residential environment have had a considerable impact on the features of postal delivery regions in recent years, resulting in a large increase in the overall workload of domestic postal delivery services. In this paper, we provide complex analysis results for postal delivery areas using various unsupervised learning approaches. First, we extract highly influential features using several feature-engineering methods. Then, using quantitative and qualitative cluster analyses, we find the distinctive traits and semantics of postal delivery zones. Unsupervised learning approaches are useful for successfully grouping postal service zones, according to our findings. Furthermore, by comparing a postal delivery region to other areas in the same group, workload balancing was achieved.

“We don’t all have the privilege of having our own quiet place”: College Students in a Hispanic Serving Institution during the COVID-19 Pandemic

ABSTRACT The COVID-19 pandemic has impacted nearly all aspects of daily life, including academic experiences for college students. While studies have highlighted the various barriers to virtual learning and with many institutions adopting a mix of synchronous and asynchronous options, there is limited evidence of what is ideal for students dealing with the ongoing challenges of the pandemic. This exploratory study uses mixed methods to assess the perception of college students on the type of virtual learning, their opportunities for engagement and expectations of post pandemic education at a Hispanic Serving Institution. The findings highlight emergent themes including resources, communication, empathy, workload, and academic-life balance. Students’ voices recommend preference for hybrid as the most popular, synchronous courses over asynchronous, and suggest students’ sense of empathy for faculty, their workload substantially increased than face-to-face instruction

Meta-Heuristic Algorithms for Multi-Objective Home Health Care Routing and Scheduling Problem Considering Time Windows and Workload Balance of Nurses

Meta-Heuristic Algorithms for Multi-Objective Home Health Care Routing and Scheduling Problem Considering Time Windows and Workload Balance of Nurses

Workforce scheduling incorporating worker skills and ergonomic constraints

In the last few decades, studies have demonstrated the correlation between worker well-being and the performance of production systems. This paper addresses the problem of assigning workers to tasks in a workshop system. In this context, recent researches have focused on the ergonomics assessment, often neglecting the evaluation of the workers’ performance. This study aims to formulate a mixed integer linear programming model to solve the workforce scheduling problem and improve the performance of the system integrating ergonomics and human skills. To overcome the complexity of the combinatorial problem, a constructive heuristic procedure is developed. Moreover, a novel approach is proposed to determine the workers’ skills. Human performance is modelled in terms of the time required to perform consecutive tasks, considering different sequences of tasks. In addition, the model was applied to a real case study to verify its feasibility. Different scenarios are tested, considering different levels of exposure to different risk factors. The results indicate that a limited increase in the makespan enables decreasing the risk level and the achievement of an excellent workload balance among workers in terms of time spent in performing tasks. Moreover, the heuristic procedure has demonstrated to perform well on instances of realistic size, and it could be adapted to many manufacturing systems to solve the problem in real industrial contexts.

Multi-Period Workload Balancing in Last-Mile Urban Delivery

In the daily dispatching of last-mile urban delivery, a delivery manager has to consider workload balance among couriers to maintain workforce morale. We consider two types of workload: incentive workload, which relates to the delivery quantity and affects a courier’s income, and effort workload, which relates to the delivery time and affects a courier’s health. Incentive workload has to be balanced over a relatively long period of time (a payroll cycle—a week or a month), whereas effort workload has to be balanced over a relatively short period of time (a shift or a day). We introduce a multi-period workload balancing problem under stochastic demand and dynamic daily dispatching, formulate it as a Markov decision process (MDP), and derive a lower bound on the optimal value of the MDP model. We propose a balanced penalty policy based on cost function approximation and use a hybrid algorithm combining the modified nested partitions method and the KN++ procedure to search for an optimal policy parameter. A comprehensive numerical study demonstrates that the proposed balanced penalty policy performs close to optimal on small instances and outperforms four benchmark policies on large instances, and provides insight into the impact of demand variation and a manager’s importance weighting of operating cost and workload balance.

A divide and combine method for machine configuration and workload balancing problem in multiple product PCB assembly

A divide and combine method for machine configuration and workload balancing problem in multiple product PCB assembly

Assembly line balancing and activity scheduling for customised products manufacturing

Nowadays, end customers require personalized products to match their specific needs. Thus, production systems must be extremely flexible. Companies typically exploit assembly lines to manufacture produces in great volumes. The development of assembly lines distinguished by mixed or multi models increases their flexibility concerning the number of product variants able to be manufactured. However, few scientific contributions deal with customizable products, i.e., produces which can be designed and ordered requiring or not a large set of available accessories.This manuscript proposes an original two-step procedure to deal with the multi-manned assembly lines for customized product manufacturing. The first step of the procedure groups the accessories together in clusters according to a specific similarity index. The accessories belonging to a cluster are typically requested together by customers and necessitate a significant mounting time. Thus, this procedure aims to split accessories belonging to the same cluster to different assembly operators avoiding their overloads.The second procedure step consists of an innovative optimization model which defines tasks and accessory assignment to operators. Furthermore, the developed model defines the activity time schedule in compliance with the task precedencies maximizing the operator workload balance. An industrial case study is adopted to test and validate the proposed procedure. The obtained results suggest superior balancing of such assembly lines, with an average worker utilization rate greater than 90%. Furthermore, in the worst case scenario in terms of customer accessories requirement, just 4 line operators out of 16 are distinguished by a maximum workload greater than the cycle time.

A Scalable Quantum Key Distribution Network Testbed Using Parallel Discrete-Event Simulation

Quantum key distribution (QKD) has been promoted as a means for secure communications. Although QKD has been widely implemented in many urban fiber networks, the large-scale deployment of QKD remains challenging. Today, researchers extensively conduct simulation-based evaluations for their designs and applications of large-scale QKD networks for cost efficiency. However, the existing discrete-event simulators offer models for QKD hardware and protocols based on sequential event execution, which limits the scale of the experiments. In this work, we explore parallel simulation of QKD networks to address this issue. Our contributions lay in the exploration of QKD network characteristics to be leveraged for parallel simulation as well as the development of a parallel simulation framework for QKD networks. We also investigate three techniques to improve the simulation performance including (1) a ladder queue based event list, (2) memoization for computationally intensive quantum state transformation information, and (3) optimization of the network partition scheme for workload balance. The experimental results show that our parallel simulator is 10 times faster than a sequential simulator when simulating a 128-node QKD network. Our linear-regression-based network partition scheme can further accelerate the simulation experiments up to two times over using a randomized network partition scheme.

Task Scheduling in Cloud Using ACO

Background: Cloud computing is a multi-tenant model for computation that offers various features for computing and storage based on user demand. With increasing cloud users, the usage increases that highlights the problem of load balancing with limited resource availability based on dynamic cloud environment. In such cases, task scheduling creates fundamental issue in cloud environment. Introduction: Certain problems such as, inefficiencies in load balancing latency, throughput ratio, proper utilization of the cloud resources, better energy consumption and response time have been observed. These drawbacks can be efficiently resolved through the incorporation of efficient load balancing and task scheduling strategies. Method: In this paper, we develop an efficient co-operative method to solve the most recent approaches against load balancing and task scheduling have been proposed using Ant Colony Optimization (ACO). These approaches enables in the clear cut identification of the problems associated with the load balancing and task scheduling strategies in the cloud environment. Results: The simulation is conducted to find the efficacy of the improved ACO system for load balancing in cloud than the other methods. The result shows that the proposed method obtains reduced execution time, reduced cost and delay. Conclusion: A unique strategic approach is developed in this paper, Load Balancing, which works with the ACO in relation to the cloud workload balancing task through the incorporation of the ACO technique. The strategy for determining the applicant nodes is based on which the load balancing approach would essentially depend. By incorporating two different approaches: the maximum minute rules and the forward-backward ant, this reliability task can be established. This method is intended to articulate the initialization of the pheromone and thus upgrade the relevant cloud-based physical properties.

Analisis Kinerja Petugas Kesehatan Dalam Penemuan Kasus Baru Stunting Pada Balita Diwilayah Kerja Dinas Kesehatan Kabupaten Ogan Komering Ulu Tahun 2021

Stunting prevention efforts are a priority in stunting prevention to reduce the incidence of stunting and prevent the impact. The role of health workers is no less important in preventing stunting. This study aims to analyze the factors related to the performance of health workers in finding new cases of stunting in Ogan Komering Ulu Regency in 2021. This study is a quantitative study, using a cross sectional design. The sample in this study was a total population of 175 respondents. Bivariate analysis used in this study was chi square and multivariate analysis used multiple logistic regression. The results showed that there was a relationship between training (p value 0.000), workload (p value 0.031), leadership (p value 0.000) and infrastructure (p value 0.012) with the performance of health workers in finding new cases of stunting in children under five in the Work Area. Ogan Komering Ulu District Health Office in 2021. Meanwhile, education (p value 0.450), years of service (p value 0.120), attitudes (p value 0.602) and motivation (p value 0.636) were not related to the performance of health workers in finding new cases of stunting in children under five in the Ogan Komering Uu District Health Office Work Area in 2021. The most dominant variable affecting the performance of health workers in finding new cases of stunting in toddlers in the Ogan Komering Ulu District Health Service Work Area in 2021 is leadership (OR 3,954). Based on the results of the study, it is suggested to the Health Office of Ogan Komering Ulu Regency to further increase training activities for health workers, complete supporting facilities and infrastructure, propose rewards for village midwives and nutrition workers, provide a balanced workload and add officers with different backgrounds. nutrition education.

An Edge Server Placement Method Based on Reinforcement Learning.

In mobile edge computing systems, the edge server placement problem is mainly tackled as a multi-objective optimization problem and solved with mixed integer programming, heuristic or meta-heuristic algorithms, etc. These methods, however, have profound defect implications such as poor scalability, local optimal solutions, and parameter tuning difficulties. To overcome these defects, we propose a novel edge server placement algorithm based on deep q-network and reinforcement learning, dubbed DQN-ESPA, which can achieve optimal placements without relying on previous placement experience. In DQN-ESPA, the edge server placement problem is modeled as a Markov decision process, which is formalized with the state space, action space and reward function, and it is subsequently solved using a reinforcement learning algorithm. Experimental results using real datasets from Shanghai Telecom show that DQN-ESPA outperforms state-of-the-art algorithms such as simulated annealing placement algorithm (SAPA), Top-K placement algorithm (TKPA), K-Means placement algorithm (KMPA), and random placement algorithm (RPA). In particular, with a comprehensive consideration of access delay and workload balance, DQN-ESPA achieves up to 13.40% and 15.54% better placement performance for 100 and 300 edge servers respectively.

Deep reinforcement learning-based multi-objective edge server placement in Internet of Vehicles

In the typical scenario of the Internet of Vehicles (IoV), the edge servers (ESs) are laid out near the road side units (RSUs) to process the collected data for a variety of IoV services in real time. Generally, because ESs are lightweight compared with cloud servers, if the ESs are not appropriately distributed, it will cause the unbalanced workload of the ESs. Thus, developing an ES plan to avoid the risk of overload and improve the quality of service (QoS) remains a challenge. To tackle it, a deep reinforcement learning-based multi-objective edge server placement strategy, named DESP, is fully explored, to promote the coverage rate, the workload balancing and reduce the average delay of finishing tasks in the IoV. In particular, the Markov Decision Process (MDP) of the ES placement problem is formulated and the deep reinforcement learning, i.e., Deep Q-Network (DQN) is applied to obtain the optimal placement scheme achieving the multiple objectives above. At last, a real vehicular data set is used for assessing the validity of DESP.

A goal programming model for the optimization of log logistics considering sorting decisions and social objective

Log logistics include sorting, processing, and transporting of logs from their place of harvest to demand locations. These activities account for a significant portion of the total log procurement costs; therefore, attempts were made in previous studies to optimize some aspects of log logistics. However, operational details, such as sorting decisions, truck compatibility requirements, and social objectives, are often disregarded in the optimization literature. Incorporating these details into the model makes the results more realistic and applicable. To address these gaps, a bi-objective mixed-integer programming model is developed in this paper to optimize log logistics. The first objective is to minimize total logistics costs, and the second objective is to provide a balanced workload for trucking contractors. The bi-objective model is solved using the goal programming approach. The model is applied to log logistics of a large Canadian forest company, where trucking contractors use heterogeneous fleet of trucks to carry various log sorts from cutblocks to sort yards for sorting. The planning horizon is 4 weeks with daily decisions. The goal programming model generates balanced workloads for the contractors with less than 0.4% increase in total costs compared to the single objective model where only the total cost is minimized.

Fault-tolerant scheduling and data placement for scientific workflow processing in geo-distributed clouds

Scientific workflow processing in geo-distributed cloud is crucial for the scheduling of large-scale tasks and the massive data placement among tasks. However, the task execution time and energy consumption of data transmission are two urgent issues when a scientific workflow is processed in the different geo-distributed data centers. Aiming at the data placement problem, this paper proposes a Lagrangian relaxation method. This method considers the workload balance, storage capacity, data dependency, transmission bandwidth, and transmission cost to obtain the minimum time of data transmission time. Aiming at the task scheduling problems, a fault-tolerant scheduling strategy is proposed. The strategy optimizes the task scheduling mechanism by considering the task execution time and energy consumption. Finally, the performance of the proposed methods is evaluated via extensive experiments In terms of the data placement, the experiment results imply that the data transmission time of the proposed relaxation algorithm can averagely achieve up to 14.61%, 38.03%, and 39.57% reduction of over that of ILP-FDP algorithm, GA-DPSO algorithm, and GPDP algorithm, respectively. As for the fault-tolerant scheduling, the energy consumption of the TSPT algorithm is the lowest. Compared with the MTS algorithm, EODS algorithm and EWTS algorithm, the average gains of the proposed algorithm are 15.33%, 16.65%, and 28.96%, respectively. Compared with the benchmark algorithms, the task execution time of the TSPT algorithm can averagely reduce up to 12.78,18.85 and 25.65, respectively.

Distributed Triangle Approximately Counting Algorithms in Simple Graph Stream

Recently, the counting algorithm of local topology structures, such as triangles, has been widely used in social network analysis, recommendation systems, user portraits and other fields. At present, the problem of counting global and local triangles in a graph stream has been widely studied, and numerous triangle counting steaming algorithms have emerged. To improve the throughput and scalability of streaming algorithms, many researches of distributed streaming algorithms on multiple machines are studied. In this article, we first propose a framework of distributed streaming algorithm based on the Master-Worker-Aggregator architecture. The two core parts of this framework are an edge distribution strategy, which plays a key role to affect the performance, including the communication overhead and workload balance, and aggregation method, which is critical to obtain the unbiased estimations of the global and local triangle counts in a graph stream. Then, we extend the state-of-the-art centralized algorithm TRIÈST into four distributed algorithms under our framework. Compared to their competitors, experimental results show that DVHT-i is excellent in accuracy and speed, performing better than the best existing distributed streaming algorithm. DEHT-b is the fastest algorithm and has the least communication overhead. What’s more, it almost achieves absolute workload balance.

Optimization of long-term planning with a constraint satisfaction problem algorithm with a machine learning

The object of the long-term planning in shipyards is to assign the ordered vessels to the berths with the consideration of the workload balancing. However, there are limitations in establishing an optimized long-term plan because the workload balancing takes too much time due to the size and the complexity of the problem domain. Most shipyards currently overcome the limitations by dividing the long-term planning into two-phase of the berth planning and the capacity planning. The berth planning is being conducted with a heuristic method by considering some rules such as the berth priority and the closeness to delivery date. Then it is followed by the capacity planning, in which the workload data is considered for the workload balancing with the previously planned data. However, the heuristic method has a fundamental problems that the optimized solution is not guaranteed owing to the limits of the search range. Also, the previous production record cannot reflect the newly ordered vessel's workload precisely. In this study, a constraint satisfaction technique is used for the optimization of the berth planning. In addition, the workload prediction model is developed based on the supervised learning with a deep neural network. Finally, proposed methods are tested with the shipyard actual data, that shows the improved results.

The Role Of Work Stress In Mediating The Influence Of Workload And Work Balance On Employee Performance of The Regional Development Planning Agency Of Aceh Province During The Covid-19 Pandemic

This study examines the work stress role in mediating the effect of workload and work balance on employee performance of the Regional Development Planning Agency of Aceh Province (Bappeda Aceh) during the covid-19 pandemic. The population was all employees at Bappeda Aceh, totaling 209 people. The sampling technique uses the census method. The model was tested using Structural Model by AMOS application. The survey was carried out during the COVID-19 pandemic, which was still hitting the end of 2021 to early 2022. The results shows that when covid 19 occurred, workload positively affects the work stress of Bappeda Aceh employees, work balance negatively affects the work stress of Bappeda Aceh employees, workload negatively affects the Bappeda Aceh employee performance, work balance positively affects the Bappeda Aceh employee performance, work stress negatively affects the Bappeda Aceh employee performance, work stress partially mediates the workload effect on the Bappeda Aceh employee performance, and work stress partially mediates the work balance effect on the Bappeda Aceh employee performance. From these findings, it is explained that the model for improving the Bappeda Aceh employee performance during the COVID-19 pandemic is a function of reducing workloads and increasing work balance so that they can reduce employee work stress. In other words, a decrease in workload during the COVID-19 period can play a role in reducing work stress so that it can improve performance, and improving work balance during the COVID-19 period, it can also reduce work stress and will have an impact on improving performance.

A Review on Collaborative Robot Assembly Line Balancing Problems

Assembly line balancing problems have been the subject of numerous studies for decades. The recent advantages of technologies in the Industry 4.0 era and integrating collaborative robots into the assembly systems have created task allocation, workload balance, and scheduling challenges. This paper provides a brief literature review on the existing collaborative assembly line balancing problems comparing their main characteristics and suggests some research directions for future studies.

Multi-Objective Task Scheduling Optimization for Load Balancing in Cloud Computing Environment Using Hybrid Artificial Bee Colony Algorithm With Reinforcement Learning

Workload balancing in cloud computing is still challenging problem, especially in Infrastructure as a Service (IaaS) in the cloud model. A problem that should not occur during cloud access is a host or server being overloaded or underloaded, which may affect the processing time or may result in a system crash. Therefore, to prevent these problems, an appropriate schedule of access should be considered so that the system can distribute tasks across all available resources, which is called load balancing. The load balancing technique should ensure that all Virtual Machines (VMs) are used appropriately. In this paper, an independent task scheduling approach in cloud computing is proposed using a Multi-objective task scheduling optimization based on the Artificial Bee Colony Algorithm (ABC) with a Q-learning algorithm,which is a reinforcement learning technique that helps the ABC algorithm work faster, called the MOABCQ method. The proposed method aims to optimize scheduling and resource utilization, maximize VM throughput, and create load balancing between VMs based on makespan, cost, and resource utilization, which are limitations of concurrent considerations. Performance analysis of the proposed method was compared using CloudSim with the existing load balancing and scheduling algorithms: Max-Min, FCFS, HABC_LJF, Q-learning, MOPSO, and MOCS algorithms in three datasets: Random, Google Cloud Jobs (GoCJ), and Synthetic workload. The experimental results indicated that the algorithms used MOABCQ approach outperformed the other algorithms in terms of reducing makespan, reducing cost, reducing degree of imbalance, increasing throughput and average resource utilization.