Maximizing Mobiles Energy Saving Through Tasks Optimal Offloading Placement in two-tier Cloud

With the massive connected devices, spectrum shortage becomes a serious problem in Internet of Things. Cognitive radio technology could mitigate this spectrum resource shortage problem in IoT. In this paper, we study the joint spectrum sensing, bandwidth allocation and energy consumption problem in cognitive IoT. We balance the tradeoff between bandwidth allocation and energy consumption, and meanwhile maximize the utility of each secondary IoT device, through optimizing the spectrum sensing period and sensing performance. We formulate the problem as a non-cooperative game, and prove the existence of Nash equilibrium. We propose a False Alarm based bandwidth allocation scheme (FA-scheme). This scheme determines the sensing period according the Nash equilibrium, and allocates the spectrum bandwidth based on the false alarm performance. Simulation results show that the proposed scheme has great advantages in terms of secondary network throughput.

A representation for specifying the timing constraints of real‐time systems is presented. Four parts of descriptions in a timing constraint: the path of events, timer, types of constraints and quantities of time, are explained. Under the relations among the paths, the hierarchical and network structures can be conducted. The former can be associated with the structure resulting from applying top‐down decomposition in software development. The latter represents the execution dependence of the paths. Based upon these clear and manageable structures, the completeness and consistency of timing constraints specification can be rigorously analysed. Thus, the quality of the specification of timing constraints can be improved during software development. A frame‐like software requirements specification language (FSRSL) is extended to specify timing constraints and the related descriptions in the paper.

This paper describes two methods to specify timing constraints in behavioral VHDL for high-level synthesis purposes. The first method specifies timing constraints on sequences of statements by using predefined procedures. The second method provides support for specification of timing constraints across process borders based on concurrent assert statements on signal events. The paper discusses also an approach to synthesize hardware with timing constraints and concentrates in particular on how to ensure consistency between the behavior of the simulation model and that of the synthesized hardware.

With the popularity of mobile devices such as smartphones and tablets, the improvement of service of quality is an important issue facing great challenges. The improvement of user service of quality is mainly reflected in reducing the energy consumption of mobile devices and the delay of task execution. Multi-access edge computing sinks computing and storage capabilities from the remote cloud to the edge network, which can effectively reduce the high latency caused by the transmission of tasks between the mobile device and the remote cloud and the high energy consumption of tasks performed locally. Most of the previous work was limited to service of quality optimization through dynamic service layout, while ignoring the critical impact of access network selection on network congestion. This article studies the task offloading model of multiple tasks and services with several MEC servers, and jointly optimizes the MEC's access network selection and service placement issues. Considering the delay and energy consumption caused by task offloading and execution, this article designs an effect function on delay and energy consumption, and aims to minimize this function to solve the MEC problem. Since this problem is NP-hard, this article designs a new optimization algorithm based on particle swarm optimization to solve this problem. Extensive simulation experiments show that the proposed optimization algorithm realize better performance than other algorithms. The algorithm has achieved good results in terms of time delay and energy consumption, which effectively reduces the system cost.

Deceptive path planning (DPP) aims to find a path that minimizes the probability of the observer identifying the real goal of the observed before it reaches. It is important for addressing issues such as public safety, strategic path planning, and logistics route privacy protection. Existing traditional methods often rely on “dissimulation”—hiding the truth—to obscure paths while ignoring the time constraints. Building upon the theory of probabilistic goal recognition based on cost difference, we proposed a DPP method, DPP_Q, based on count-based Q-learning for solving the DPP problems in discrete path-planning domains under specific time constraints. Furthermore, to extend this method to continuous domains, we proposed a new model of probabilistic goal recognition called the Approximate Goal Recognition Model (AGRM) and verified its feasibility in discrete path-planning domains. Finally, we also proposed a DPP method based on proximal policy optimization for continuous path-planning domains under specific time constraints called DPP_PPO. DPP methods like DPP_Q and DPP_PPO are types of research that have not yet been explored in the field of path planning. Experimental results show that, in discrete domains, compared to traditional methods, DPP_Q exhibits better effectiveness in enhancing the average deceptiveness of paths. (Improved on average by 12.53% compared to traditional methods). In continuous domains, DPP_PPO shows significant advantages over random walk methods. Both DPP_Q and DPP_PPO demonstrate good applicability in path-planning domains with uncomplicated obstacles.

We consider the problem of minimizing the continuous valued total variation subject to different unary terms on trees and propose fast direct algorithms based on dynamic programming to solve these problems. We treat both the convex and the non-convex case and derive worst case complexities that are equal or better than existing methods. We show applications to total variation based 2D image processing and computer vision problems based on a Lagrangian decomposition approach. The resulting algorithms are very efficient, offer a high degree of parallelism and come along with memory requirements which are only in the order of the number of image pixels.

Microalgae are a promising biomass for the biobased economy to produce food, feed, fuel, chemicals and materials. So far, large-scale production of algae is limited and as a result estimates on the performance of such large systems are scarce. There is a need to estimate large-scale biomass productivity and energy consumption, while considering the uncertainty and complexity in such large-scale systems. In this thesis frameworks are developed to assess 1) the productivity during algae cultivation, 2) energy consumption during the transport of resources and processing biomass to biodiesel, and 3) the frameworks are applied to estimate the impact of algae cultivation in the production of algae-based food commodities. Design, location and future scenario are applied to deal with the complexity and uncertainty arising in the various data and models used. The first part of this thesis focuses on the development of a productivity framework for biomass production for flat panels (Chapter 2), horizontal and vertical tubular photobioreactors (Chapter 3) and raceway ponds (Chapter 4). The framework uses bio-physics-based models to simulate the light input on the reactor surface and the light gradient inside the reactor systems. The internal light gradient depends on the reactor geometry and dimensions, and the penetration of diffuse light between parallel reactors, which includes the canyon effect, and the reflection of light from the ground surface to the reactors are incorporated as well. Specific growth rates are derived from this internal light gradient based on species-specific growth characteristics. In raceway ponds the effect of the dynamic water temperature on the specific growth rate is included. The productivity framework enables to study cultivation under a wide range of process conditions and reactor designs, even those which have not been yet developed or tested under outdoor conditions. The results show that regional weather conditions, solar angles and algae species are key factors in making the best choice for the specific reactor design. The productivity framework allows to optimise the reactor design (e.g. geometry, light path, distances between parallel units and height) to the regional light conditions and growth characteristics of the algae species of interest. The best biomass concentration for cultivation varies between the reactor design, location and algae species. We recommend to select species suited to growth well at the regional light angles and weather conditions. An initial global sensitivity analysis shows that the absorption coefficient, maximum specific growth rate and functional cross section of the photosynthetic apparatus are the essential parameters of the model for single flat panels. An important next step is to validate and calibrate the productivity framework using data from outdoor experiments in various reactor designs, at different locations and with several algae species. Algae production is strongly connected to regional weather conditions, but also to the infrastructure for resource supply and to the processing of biomass. The energy consumption for resource supply has not been quantified yet and the energy consumption of biomass processing is mostly based on fixed values. These elements are tackled in part 2 of this thesis. In Chapter 5the productivity framework is combined with logistic models to optimise the supply network for algae cultivation. The results show that the availability, supply and demand of resources has a dominant effect on the feasibility of regions for algae cultivation. Not all locations achieve a positive energy balance for transport and the supply logistics is essential for planning algae cultivation locations. In the Benelux many locations are feasible for algae production due to the availability of large amounts of resources, while the limited supply of CO2in southern France and the Sahara demands for plants which are scattered over the regions. For the Sahara the distance for water transport should be minimal. Still, the average transport distances are higher than commonly assumed and algae cultivation does not necessarily need to take place in proximity of CO2supply. The transport energy consumption is found to be low compared to the energy contained in algae biomass (mostly below 3%). Chapter 6 describesa model-based combinatorial optimisation approach for the energy-efficientprocessing of algae biomass. In this approach, mass and energy balances and additional relations are used to relate the product yield and energy consumption of process units and process routes to the processing conditions. Process routes with the highest net energy ratios are derived by optimising the process conditions of each process unit in a given superstructure. This optimisation leads to 5-38% improvement of the net energy ratio compared to fixed process conditions. The approach moreover allows a bottleneck analysis for each process route. The results show that process design should be tailor-made. The model-based approach proves to be a versatile tool for the design of efficient microalgae processing systems. The developed frameworks combined with scenario studies are a powerful tool to assess algae production. The presented approaches help to reduce the uncertainty in the interpretation of data and are thereby an appropriate basis to use in impact analysis. In Chapter 7this is illustrated for the production of algae protein and oil as food commodities. The design scenarios show the implications of various reactor designs, two algae species and at two locations on biomass productivity, production cost and environmental life cycle indicators. The achievements of this work and the new horizons from this work are discussed in Chapter 8. The results of the developed frameworks demonstrate the power of the scenario approach and show that sensible predictions and projections of biomass productivity and energy consumption for logistics and biomass processing follow from the models.

Modular design is a key challenge for enabling large-scale reuse of hardware modules. Unlike software, however, hardware designs correspond to physical circuits and inherit constraints from them. Timing constraints—which cycle a signal arrives, when an input is read—and structural constraints—how often a multiplier accepts new inputs—are fundamental to hardware interfaces. Existing hardware design languages do not provide a way to encode these constraints; a user must read documentation, build scripts, or in the worst case, a module’s implementation to understand how to use it. We present Filament, a language for modular hardware design that supports the specification and enforcement of timing and structural constraints for statically scheduled pipelines. Filament usestimeline types, which describe the intervals of clock-cycle time when a given signal is available or required. Filament enablessafe compositionof hardware modules, ensures that the resulting designs are correctly pipelined, and predictably lowers them to efficient hardware.

With an increase in the number of mobile applications, the development of mobile crowdsensing systems has recently attracted a significant attention in both academic researchers and industries. In mobile crowdsensing system, the remote cloud (or back-end server) harvests all the crowdsensing data from the mobile devices, and the crowdsensing data can be uploaded immediately via 3G/4G. To reduce the cost and energy consumption, many academic researchers and industries investigate the way of mobile data offloading. Due to the sparse distribution of the WiFi APs, the crowdsensing data is often delayed to offloading. In this paper, compared with offloading data via WiFi APs, we investigate the communication and sharing of crowdsensing data by vehicles near the event (such as a pathole on the road), termed as a local crowd. The crowd-based approach has a lower delay than the offloading-based approach, by considering the quality of truth discovery. We define an utility function related to the crowdsensing data shared by the local crowd, in order to quantify the trade-off between the quality of the truth discovery and the user satisfaction. Our extensional simulations verify the effectiveness of our proposed schemes.

Abstract
\n—Cellular network energy optimization is driven by different factors, namely transmission power and activity of the base stations. Quality of Service (QoS) of users is determined by received power, interference and bandwidth allocation and cannot be neglected. In existing studies, such drivers are treated separately, i.e., bandwidth allocation is fixed while power consumption is a variable to be optimized. This paper proposes
\na novel optimization framework which is aimed at minimizing the power consumption in cellular networks while affording a minimum bit rate for each mobile terminal by jointly considering energy consumption and QoS drivers. Mixed Integer Quadratic Programming (MIQP) based optimization framework solves the problems of the determination of the user association, the bandwidth allocation, the identification of the active base stations and their transmission power, guaranteeing also a requested service rate for each user. The proposed solution is shown to allow power savings of up to 60%, very close to the optimum lower bound, when the traffic is below the 35% of the maximum load.

Federated edge learning (FEL) deploys a machine learning algorithm by using devices distributed on the edge of a network, trains massive local data, uploads the local model to update the parameters after training, and performs alternate updating with global model parameters to reduce the pressure for uplink data transmission, prevent systematic time delay and ensure data security. This paper proposes that an optimal balance between time delay and energy consumption be achieved by optimizing the transmission power and bandwidth allocation based on user quality of experience (QoE) in a multi-server intelligent edge network. Given the limited computing capability of devices involved in FEL local training, the transmission power is modeled as a quasi-convex uplink power allocation (UPA) problem, and a lower energy consumption bandwidth allocation algorithm is proposed for solution-seeking. The proposed algorithm allocates appropriate power to the device by adapting the computing power and channel state of the device, thereby reducing energy consumption. As the theoretical deduction result suggests that additional bandwidth should be allocated to those devices with weak computing capabilities and poor channel conditions to realize minimal energy consumption within the restraint time. The simulation result indicates that, the maximum gain of the proposed algorithm can be optimized by 31% compared with the baseline.

We study the problem of providing real-time service to hard and soft real-time messages in wavelength division multiplexing (WDM) optical networks. We propose and evaluate a set of scheduling algorithms which schedule message transmissions in single-hop WDM passive star networks based on specific time constraints. We compare the performances of our algorithms with that of the typical WDM scheduling algorithm which does not consider the time constraint of the transmitted messages by discrete-event simulations and an analytical model. We find that the improvement on real-time performance can be attributed to our scheduling algorithms where the time constraint of messages has been taken into consideration. This study suggests that when scheduling real-time messages in WDM networks, one has to consider not only the problem of resources allocation in the network but also the problem of sequencing messages based on their time constraints.

Higher reliability and lower energy consumption are conflicting, yet among the most important design objectives for the real-time systems. Moreover, in the domain of real-time systems, non-preemptive scheduling is relatively unexplored with objectives such as reliability and energy. Thus we propose an active replication based framework to schedule a set of periodic real-time tasks in the non-preemptive heterogeneous environment such that the given reliability and timing constraints are satisfied whereas the energy consumption is minimized. First, we formulate the problem as a constraint optimization problem that provides an optimal solution; however, it does not scale well. Thus, we also propose heuristics which apply reservation of processors and reallocation of jobs, to compute suboptimal solution efficiently in terms of energy consumption as well as schedulability. Heuristics make use of the interplay of task-level reliability target, reliability of replicas, number of replicas, reliability of tasks, and energy consumption. We perform an experimental study on the test cases generated by extending UUnisort algorithm <xref ref-type="bibr" rid="ref1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</xref> and observe the effect of various simulation parameters on energy consumption and schedulability.

bwSlicer: A bandwidth slicing framework for cloud data centers

With the rise of artificial intelligence (AI) in recent years and the subsequent increase in complexity of the applied models, the growing demand in computational resources is starting to pose a significant challenge. The need for higher compute power is being met with increasingly more potent accelerator hardware as well as the use of large and powerful compute clusters. However, the gain in prediction accuracy from large models trained on distributed and accelerated systems ultimately comes at the price of a substantial increase in energy demand, and researchers have started questioning the environmental friendliness of such AI methods at scale. Consequently, awareness of energy efficiency plays an important role for AI model developers and hardware infrastructure operators likewise. The energy consumption of AI workloads depends both on the model implementation and the composition of the utilized hardware. Therefore, accurate measurements of the power draw of respective AI workflows on different types of compute nodes is key to algorithmic improvements and the design of future compute clusters and hardware. Towards this end, we present measurements of the energy consumption of two typical applications of deep learning models on different types of heterogeneous compute nodes. Our results indicate that 1. contrary to common approaches, deriving energy consumption directly from runtime is not accurate, but the consumption of the compute node needs to be considered regarding its composition; 2. neglecting accelerator hardware on mixed nodes results in overproportional inefficiency regarding energy consumption; 3. energy consumption of model training and inference should be considered separately – while training on GPUs outperforms all other node types regarding both runtime and energy consumption, inference on CPU nodes can be comparably efficient. One advantage of our approach is the fact that the information on energy consumption is available to all users of the supercomputer and not just those with administrator rights, enabling an easy transfer to other workloads alongside a raise in user-awareness of energy consumption.KeywordsEnergy measurementArtificial intelligenceGreen AIEnergy efficiencyHigh performance computingGPUs