Showing papers on "Resource management published in 2022"

Decentralized Edge Intelligence: A Dynamic Resource Allocation Framework for Hierarchical Federated Learning

Abstract: To enable the large scale and efficient deployment of Artificial Intelligence (AI), the confluence of AI and Edge Computing has given rise to Edge Intelligence, which leverages on the computation and communication capabilities of end devices and edge servers to process data closer to where it is produced. One of the enabling technologies of Edge Intelligence is the privacy preserving machine learning paradigm known as Federated Learning (FL), which enables data owners to conduct model training without having to transmit their raw data to third-party servers. However, the FL network is envisioned to involve thousands of heterogeneous distributed devices. As a result, communication inefficiency remains a key bottleneck. To reduce node failures and device dropouts, the Hierarchical Federated Learning (HFL) framework has been proposed whereby cluster heads are designated to support the data owners through intermediate model aggregation. This decentralized learning approach reduces the reliance on a central controller, e.g., the model owner. However, the issues of resource allocation and incentive design are not well-studied in the HFL framework. In this article, we consider a two-level resource allocation and incentive mechanism design problem. In the lower level, the cluster heads offer rewards in exchange for the data owners’ participation, and the data owners are free to choose which cluster to join. Specifically, we apply the evolutionary game theory to model the dynamics of the cluster selection process. In the upper level, each cluster head can choose to serve a model owner, whereas the model owners have to compete amongst each other for the services of the cluster heads. As such, we propose a deep learning based auction mechanism to derive the valuation of each cluster head's services. The performance evaluation shows the uniqueness and stability of our proposed evolutionary game, as well as the revenue maximizing properties of the deep learning based auction.

Intelligent-Driven Green Resource Allocation for Industrial Internet of Things in 5G Heterogeneous Networks

Abstract: The Industrial Internet of Things (IIoT) is one of the important applications under the 5G massive machine type of communication (mMTC) scenario. To ensure the high reliability of IIoT services, it is necessary to apply an efficient resource allocation method under the dynamic and complex environment. In view of the absence of energy-efficient resource management architecture for the entire network, this article proposes an intelligent-driven green resource allocation mechanism for the IIoT under 5G heterogeneous networks. First, an intelligent end-to-end self-organizing resource allocation framework for IIoT service is given. Next, an energy-efficient resource allocation model within the framework is proposed. It is then solved by an intelligent mechanism with the asynchronous advantage actor critic driven deep reinforcement learning algorithm. Through the comparison analysis of different methods and rewards under IIoT scenarios with proper parameters setting, the proposed method can achieve better performance than other traditional deep learning (DL) methods and maintain service quality above accepted levels as well.

Resource Management Framework Using Deep Neural Networks in Multi-Cloud Environment

Abstract: In recent decades, there exists a wide increase in traffic of clouds due to the enormous increase of media content. The popularity of cloud is gaining its attention due to its ease of use and flexible model but suffers from poor resource2 management and its minimal extendibility of service portfolio. However, with recent advancement, the services are effectively managed, and its discovery is made further possible. To handle larger amounts of multimedia contents in a standalone cloud, deployment of wide operable systems is yet required that handles the data effectively with increasing demands of the user. A resource allocation framework is designed in this paper that uses a gray wolf optimization (GWO) architecture to effectively learn the operation of resource allocation in an optimal manner. For optimal service provisioning and scalability, the cloud at times communicates with each other based on the resource allocated by the deep neural network, and then the resources are shared. Such a scenario forms the multi-cloud computing, and the resource management using the deep neural network ensures trivial solutions on poor scalability. The deep neural network acts as a model for controlling the routing capabilities based on the input data rate and the storage space available in the multi-clouds. The deep neural network operates in such a way that it reduces the delay in processing and storage of data to cloud that ensures flexible operations across the cloud. The entire operation is divided into two modules: the first module operates on data processing and routing operations, and the second module acts as a control plane using the deep neural network that ensures optimal allocation of resources based on the data obtained and processed in the first module. These two models ensure better delivery of data to the cloud with proper allocation and storage of resources in the multi-cloud environment. The simulation is conducted using Java (netbeans) platform, and it is evaluated further using CloudSim toolkit. The results are experimented on various performance metrics that includes time delay and cost of resource allocation on multi-cloud.

Linkage of impact pathways to cultural perspectives to account for multiple aspects of mineral resource use in life cycle assessment

Abstract: Important advances have been made to define the multiple impact pathways relating mineral resource use to the area of protection (AoP) natural resources in life cycle assessment (LCA). Yet, the link between stakeholders’ interests and the aspects relevant to resource use as addressed by existing impact assessment methods has so far only marginally been explored. This article proposes to go beyond the case-specific determination of stakeholders’ interests (and the associated selection of impact assessment method) by defining multiple groups of different values based on cultural perspectives, in order to determine the corresponding relevant impact pathways and assessment methods. Relying on the Cultural Theory and related potential development scenarios, we identify socio-economic objectives and resource management strategies that fit the egalitarian, individualist and hierarchist perspectives. Our analysis reveals that different aspects of resource use may be most relevant to assess for each perspective since they pursue different socio-economic objectives. Egalitarians are expected to prioritize the long-term availability of geological stocks for future generations by keeping extraction flows to a minimum to reach global sufficiency, and individualists, to safeguard their short-term accessibility to resources by managing their supply risk. Hierarchists are likely to aim to maximize the value obtained from resources globally, and could thus focus on addressing dissipative flows. Building on this analysis, we provide a proposal for a more holistic assessment of the impact pathways linked to mineral resource use using existing LCIA methods, and identify ways forward for method developments to come.

Centralised Quality of Experience and Service Framework Using PROMETHEE-II for Cloud Provider Selection

Abstract: The extensive diffusion of cloud services has fostered a business growth culture and innovation that propagate to many consumers and providers. For enabling a sustainable trusted relationship and for forming practicable successful service level agreements (SLAs), all stakeholders need a centralised Quality of Experience (QoE) and Quality of Service (QoS) repository that assists them in forming such an agreement. A cloud consumer needs a centralised QoE repository that supports them in selecting the right service provider that satisfies consumer’s requirements in terms of cost, reliability, efficiency and other QoS parameters. On the other end, a cloud provider needs a reliable QoS repository that provides consumers with up-to-date information about services and enables a provider to take an optimal decision to allocate the amount of marginal resources while forming an SLA. Due to the elastic nature of a cloud and lack of proper resource management, the service provider usually caught in service violation, leading to violation penalties both in terms of trust and money. Existing literature lacks studies on a centralised repository to assist cloud providers in resource management and cloud consumer service selection. To address the issue, we discuss the idea of a Centralised Quality of Experience and Service (CQoES) repository framework. The approach uses PROMETHEE-II method where each alternatives are assessed based on consumer’s custom weighted QoS attributes. The framework ensures the cloud marketplace’s economic growth and helps the interacting parties build a durable and long-term trusted relationship.

An integrated deep learning and stochastic optimization approach for resource management in team-based healthcare systems

Abstract: The aging of the global population and the increasing number of patients with chronic diseases necessitate an efficient healthcare operations mechanism to enable provision of appropriate services to patients in a timely and cost-efficient manner. This research provides a solution for two unanswered and critical challenges in healthcare team-based resource planning by employing machine learning and stochastic optimization. The first challenge is how the required workload of a patient should be measured and predicted. The second challenge is how decision-makers should plan and optimize resources in a healthcare team and eventually allocate patients to the available resources to efficiently satisfy needs and minimize costs. In this research, we develop a novel integrated model that provides a mathematical and systematic solution for predicting healthcare providers' total workload and balancing their workload when the required workload is unknown. The proposed approach consists of predictive and prescriptive phases. First, we predict the required workload for different patient types by proposing a deep multi-task learning approach. Then, we use the result of the prediction stage as input for assigning every patient to one of the available healthcare teams, determining the number of required teams, and balancing the teams' workloads in the prescriptive decision-making stage. The outcome of this study suggests that using multi-task learning on represented data outperforms other conventional prediction methods. Moreover, the results of using the proposed stochastic optimization model for resource planning indicate that consideration of randomness and stochastic variables in modeling team-based resource allocation reduces the total cost of healthcare operations considerably, and as a result, leads to enhanced access to healthcare.

Comparison of Machine Learning Techniques for VNF Resource Requirements Prediction in NFV

Abstract: The network function virtualization (NFV) is a developing architecture that uses virtualization technology to separate software from hardware. One of the most important challenges of NFV is the resource management of virtualized network functions (VNFs). According to the dynamic nature of the NFV, resource requirements of VNFs do not always remain static. In fact, the resource allocation to VNFs must be changed to correspond to variations of incoming traffic to the network. These changes cause a significant delay in the reallocation of resources. For this reason, applying resource estimation models before their allocation can prevent the upcoming problems and leads to performance improvement of resource allocation methods dynamically. In this paper, according to the resource prediction importance in NFV, three support vector regression (SVR), decision tree (DT) and k-nearest neighbor (KNN) algorithms of machine learning techniques are analyzed and compared. In addition, the effect of the genetic algorithm as a feature selection method on the mentioned methods is evaluated. The results show that an error less than one in SVR, DT, and KNN algorithms in predicting resources is achieved. However, the SVR algorithm has more execution time than the other two algorithms.

Smart Energy Management: A Computational Approach

Abstract: Among the practitioners in the energy management domain, there is enormous excitement about synthesizing and benefiting from numerous technologies, including real-time monitoring, net metering, demand response, distributed generation from intermittent sources such as solar and wind, active control of power flows, enhanced storage capabilities, and micro-grids. A common theme in today’s solutions is the data-driven nature of the enabling technologies — to analyze requirements, use measurement/monitoring data to drive actuation/control, optimization, and resource management. The ability of modern sensing and IOT (Internet of Things) devices to inform us about the current state of the system and provide a timely and state-appropriate (rather than a broad, imprecise) response, backed up by analysis leads to novel solutions that are also practical and efficient.

Beyond energy services: A multidimensional and cross-disciplinary agenda for home energy management research

Abstract: Home Energy Management (HEM) has a significantly growing impact on strategic energy policy, digital equity, as well as housing development and transport issues. With the proliferation of home working, reliance on electricity for heating and cooling and the increasing needs for electric charging for transportation, there is an urgent need to develop novel ways for efficient management of home energy use. Current efforts focus on HEM technologies at individual household levels, without considering the social or spatial context or their collective community-wide interrelated dependencies. We propose a multifaceted agenda at the intersection of disciplinary domains to tackle this problem by using a multidimensional lens that draws on energy behaviour, architectural research, biomimetics, and computational design, simultaneously. Optimal and effective behavioural patterns can be extracted and abstracted from nature, informing a more collective and interrelated behavioural dependencies approach that considers the complex multidimensional energy use patterns of different housing typologies. This paper discusses the analytical benefits of this new research approach through a study of home energy management behaviour. The approach though could be expanded to consider other similar empirical contexts whereby sustainable multidimensional resource management is sought such as water use, food distribution as well as transport and mobility.

MQTT Based Appliance Control and Automation with Room Occupancy Monitoring Using YOLO

Abstract: Control and automation systems are desired in today’s fast-paced world as they provide users the luxury of accessing and controlling devices remotely and automate their activities to an extent. This eventually produces desired results such as efficient energy and resource management. In this study a control and automation system was developed, specifically a smart air conditioning system guided by room occupancy accounting. The typical architecture of a home automation system was used as the system’s architecture, wherein different inherent and non-inherently smart devices were integrated through the Internet IoT platform OpenHAB and MQTT as communication protocol. The development of this study served as demonstration of control and automation systems’ capability to be scaled to larger paradigms such as smart buildings and facilitate efficient energy and resource management on a larger scale.

Water-Energy-Food Nexus Review for Biofuels Assessment

Abstract: The appropriate use of limited natural resources for generating basic human needs such as energy, food, and water, is essential to help the society function efficiently. Hence, a new approach called nexus is being considered to resolve the effects of intrinsic trade-offs between the essential needs. A review of different methods and frameworks of the water-energy-food nexus was done in this article to give a detailed repository of information on existing approaches and advocate the development of a more holistic quantitative nexus method. Assessing biofuels under the water-energy-food nexus perspective, this review addresses the sustainability of bioenergy production. The results show the countries that can sustainably produce first-generation biofuels. Only a few methods have varied interdisciplinary procedures to analyse the nexus, and more analytical software and data on resource availability/use are needed to address trade-offs between these interacting resource sectors constituting the nexus. Also, “land” is suggested as an additional sector to consider in future studies using both the nexus index and life cycle assessment methodology. The review reveals that to tackle composite challenges related to resource management, cross-disciplinary methods are essential to integrate environmental, socio-political facets of water, energy, and food; employ collaborative frameworks; and seek the engagement of decision-makers.

Fog–Cloud-Assisted Internet of Things: A Review of Workload Allocation and Latency Management Techniques

Abstract: The vast amount of data generated from the various Internet of things (IoT) applications and management of such applications has become a major concern for researchers. Cloud computing can manage these situations but the distance between such applications and cloud data centres create havoc when latency is concerned. For handling such scenarios, where cloud alone cannot handle latency sensitive and real-time data analytics, the role of fog computing comes into the picture. Fog computing works in between the cloud computing and the IoT applications. Working as an intermediary it provides resource management, infrastructure monitoring, and data management. Sensors and actuators provide additional monitoring components for IoT applications like health care, surveillance, etc. This paper discusses the problems with the existing cloud infrastructure as far as IoT application deployment is concerned and how fog computing assists IoT applications for the smooth running. The recent developments specifically related to workload allocation and latency management are the highlight of this paper.

Resource Management Challenges in IoT Based Healthcare System

Abstract: Internet of thing based healthcare (IoThc) system is a platform which is an integration of smarter device, intelligent processor and informative communication. It also requires a huge enhancement to compete with the current scenario and medication facilities. IoT has a huge architecture that clarifies the related field. Still, some lack of knowledge is opposing the principal investigators to gain more information about it. The proliferation of IoT and its healthcare applications is increasing continuously. These applications like automation systems, E-Health monitoring, etc. are using heterogeneous sensors in their environment. It results in a large variety of health care data, sensing resources, and protocols which need to be managed in which will be beneficial for the treatment of patients. It also requires resource provisioning and management through the healthcare cloud. Key features of the healthcare cloud like on-demand access, resource provisioning, elasticity, etc. can be integrated with IoT domains. In this paper, IoT-based healthcare cloud with device approach is taken into consideration. It focuses on the resource management challenges by applying the cloud concept in the IoT environment. Resources are assigned to the healthcare device cloud on-demand basis. Physical devices of IoT are shared and provisioned like in the cloud environment where resources are available to serve the request using infrastructure as a service (IaaS).

Resource Management Model to Reduce Maintenance Service Times for SMEs in Lima-Peru

Abstract: Currently, companies engaged in maintenance and repair work are being affected due to the gradual increase in delays in their services. Consequently, an important economic impact has been generated in the sector, which is reflected in a 21% increase in its operating costs. After reviewing the literature, little information was found about the sector, due to which, the motivation is to carry out the study of a service SME in Lima, Peru. The proposed solution model comprises four phases: in the first phase, change management is developed, in the second phase, a study of the work is carried out, in the third phase, inventory management is applied, and in the last phase, they are standardized. The procedures previously developed. Based on this, it is possible to reduce costs by up to 12%, as well as reducing the number of delays by 15%.

Optical Wireless System Performance, Deployment, and Optimization

Abstract: As the 5G milestone approaches, there needs to be concerted effort towards practical deployment strategies and optimization in order to fully capitalize on the 5G benefits and KPIs (key performance indicators). This chapter aims to provide just that; the authors provide new insights on the performance of the optical wireless link and how this can be deployed as an integrated mobile architecture to provide small cell coverage to highly dense hotspot areas. Not only we investigate the system performance and practical deployment approaches, but we also address their optimization based on ML approaches. Virtual resource management and big data represent prominent 5G paradigms that offer new opportunities in terms of network management and optimization. In this chapter, we consider how ML can play a major role in network optimization by bringing intelligence to the limelight and by investigating how learned patterns or relationships between network states and QoE can be used towards intelligent prediction and optimization.

Resource Optimisation Problems in Cognitive Radio Networks

Abstract: Even with the implementation of the ‘best’ spectrum sensing techniques, the amount of spectrum resource that could be made available for the cognitive radio networks may still be grossly insufficient. Besides, there are other important resources such as transmission power and data rates that must be considered alongside the spectrum resource for a meaningful implementation of the cognitive radio networks. Just like the spectrum, these other resources for the cognitive radio networks are non-ubiquitous and may be insufficient to meet the demands and expectations of cognitive radio networks if not properly managed. It is therefore imperative to investigate the best approach to allocate, administer and/or manage these non-ubiquitous resources of the cognitive radio networks. This chapter discusses the resource allocation or administration problems of cognitive radio networks and establishes them as optimisation problems. A general representation of resource optimisation problems in cognitive radio networks is then provided. Some classical examples of resource allocation problems and problem formulations in cognitive radio networks are presented. Thereafter, the unique characteristics of the resource optimisation problems in cognitive radio networks that make them different from the resource problems of other communication networks are discussed.