Multi-core optimization for conjugate gradient benchmark on heterogeneous processors

Developing parallel applications on heterogeneous processors is facing the challenges of 'memory wall',due to limited capacity of local storage,limited bandwidth and long latency for memory access. Aiming at this problem,a parallelization approach was proposed with six memory optimization schemes for CG,four schemes of them aiming at all kinds of sparse matrix-vector multiplication (SPMV) operation. Conducted on IBM QS20,the parallelization approach can reach up to 21 and 133 times speedups with size A and B,respectively,compared with single power processor element. Finally,the conclusion is drawn that the peak bandwidth of memory access on Cell BE can be obtained in SPMV,simple computation is more efficient on heterogeneous processors and loop-unrolling can hide local storage access latency while executing scalar operation on SIMD cores.

Hierarchical Optimization Method for Federated Learning with Feature Alignment and Decision Fusion

In the realm of data privacy protection,federated learning aims to collaboratively train a global model.However,heterogeneous data between clients presents challenges,often resulting in slow convergence and inadequate accuracy of the global model.Utilizing shared feature representations alongside customized classifiers for individual clients emerges as a promising personalized solution.Nonetheless,previous research has frequently neglected the integration of global knowledge into local representation learning and the synergy between global and local classifiers,thereby limiting model performance.To tackle these issues,this study proposes a hierarchical optimization method for federated learning with feature alignment and the fusion of classification decisions(FedFCD).FedFCD regularizes the relationship between global and local feature representations to achieve alignment and incorporates decision information from the global classifier,facilitating the late fusion of decision outputs from both global and local classifiers.Additionally,FedFCD employs a hierarchical optimization strategy to flexibly optimize model parameters.Through experiments on the Fashion-MNIST,CIFAR-10 and CIFAR-100 datasets,we demonstrate the effectiveness and superiority of FedFCD.For instance,on the CIFAR-100 dataset,FedFCD exhibited a significant improvement in average test accuracy by 6.83%compared to four outstanding personalized federated learning approaches.Furthermore,extended experiments confirm the robustness of FedFCD across various hyperparameter values.

A Discrete Multi-Objective Squirrel Search Algorithm for Energy-Efficient Distributed Heterogeneous Permutation Flowshop with Variable Processing Speed

In the manufacturing industry,reasonable scheduling can greatly improve production efficiency,while excessive resource consumption highlights the growing significance of energy conservation in production.This paper studies the problem of energy-efficient distributed heterogeneous permutation flowshop problem with variable processing speed(DHPFSP-VPS),considering both the minimum makespan and total energy consumption(TEC)as objectives.A discrete multi-objective squirrel search algorithm(DMSSA)is proposed to solve the DHPFSPVPS.DMSSA makes four improvements based on the squirrel search algorithm.Firstly,in terms of the population initialization strategy,four hybrid initialization methods targeting different objectives are proposed to enhance the quality of initial solutions.Secondly,enhancements are made to the population hierarchy system and position updating methods of the squirrel search algorithm,making it more suitable for discrete scheduling problems.Additionally,regarding the search strategy,six local searches are designed based on problem characteristics to enhance search capability.Moreover,a dynamic predator strategy based on Q-learning is devised to effectively balance DMSSA’s capability for global exploration and local exploitation.Finally,two speed control energy-efficient strategies are designed to reduce TEC.Extensive comparative experiments are conducted in this paper to validate the effectiveness of the proposed strategies.The results of comparing DMSSA with other algorithms demonstrate its superior performance and its potential for efficient solving of the DHPFSP-VPS problem.

Stability and optimal control for delayed rumor-spreading model with nonlinear incidence over heterogeneous networks

On the multilingual online social networks of global information sharing,the wanton spread of rumors has an enormous negative impact on people's lives.Thus,it is essential to explore the rumor-spreading rules in multilingual environment and formulate corresponding control strategies to reduce the harm caused by rumor propagation.In this paper,considering the multilingual environment and intervention mechanism in the rumor-spreading process,an improved ignorants–spreaders-1–spreaders-2–removers(I2SR)rumor-spreading model with time delay and the nonlinear incidence is established in heterogeneous networks.Firstly,based on the mean-field equations corresponding to the model,the basic reproduction number is derived to ensure the existence of rumor-spreading equilibrium.Secondly,by applying Lyapunov stability theory and graph theory,the global stability of rumor-spreading equilibrium is analyzed in detail.In particular,aiming at the lowest control cost,the optimal control scheme is designed to optimize the intervention mechanism,and the optimal control conditions are derived using the Pontryagin's minimum principle.Finally,some illustrative examples are provided to verify the effectiveness of the theoretical results.The results show that optimizing the intervention mechanism can effectively reduce the densities of spreaders-1 and spreaders-2 within the expected time,which provides guiding insights for public opinion managers to control rumors.

Optimal Insurance with Background Risk under the Ambiguity and Belief Heterogeneity Structure

In this paper, we discuss the optimal insurance in the presence of background risk while the insured is ambiguity averse and there exists belief heterogeneity between the insured and the insurer. We give the optimal insurance contract when maxing the insured’s expected utility of his/her remaining wealth under the smooth ambiguity model and the heterogeneous belief form satisfying the MHR condition. We calculate the insurance premium by using generalized Wang’s premium and also introduce a series of stochastic orders proposed by [1] to describe the relationships among the insurable risk, background risk and ambiguity parameter. We obtain the deductible insurance is the optimal insurance while they meet specific dependence structures.

Cuckoo search algorithm-based optimal deployment method of heterogeneous multistatic radar for barrier coverage

This paper proposes an optimal deployment method of heterogeneous multistatic radars to construct arc barrier coverage with location restrictions.This method analyzes and proves the properties of different deployment patterns in the optimal deployment sequence.Based on these properties and considering location restrictions,it introduces an optimization model of arc barrier coverage and aims to minimize the total deployment cost of heterogeneous multistatic radars.To overcome the non-convexity of the model and the non-analytical nature of the objective function,an algorithm combining integer line programming and the cuckoo search algorithm(CSA)is proposed.The proposed algorithm can determine the number of receivers and transmitters in each optimal deployment squence to minimize the total placement cost.Simulations are conducted in different conditions to verify the effectiveness of the proposed method.

Cooperative task allocation for heterogeneous multi-UAV using multi-objective optimization algorithm

The application of multiple UAVs in complicated tasks has been widely explored in recent years.Due to the advantages of flexibility,cheapness and consistence,the performance of heterogeneous multi-UAVs with proper cooperative task allocation is superior to over the single UAV.Accordingly,several constraints should be satisfied to realize the efficient cooperation,such as special time-window,variant equipment,specified execution sequence.Hence,a proper task allocation in UAVs is the crucial point for the final success.The task allocation problem of the heterogeneous UAVs can be formulated as a multi-objective optimization problem coupled with the UAV dynamics.To this end,a multi-layer encoding strategy and a constraint scheduling method are designed to handle the critical logical and physical constraints.In addition,four optimization objectives:completion time,target reward,UAV damage,and total range,are introduced to evaluate various allocation plans.Subsequently,to efficiently solve the multi-objective optimization problem,an improved multi-objective quantum-behaved particle swarm optimization(IMOQPSO)algorithm is proposed.During this algorithm,a modified solution evaluation method is designed to guide algorithmic evolution;both the convergence and distribution of particles are considered comprehensively;and boundary solutions which may produce some special allocation plans are preserved.Moreover,adaptive parameter control and mixed update mechanism are also introduced in this algorithm.Finally,both the proposed model and algorithm are verified by simulation experiments.

SAGD Optimization for Heterogeneous Reservoir

This paper demonstrates the use of a commercial simulator as a tool with which to optimize the SAGD （steam-assisted gravity drainage） start-up phase process. The factors affecting the start-up phase are the prime targets. Among the key investigated factors are wellbore geometry effects, reservoir heterogeneity and circulation phase length. Each of the parameters was investigated via steam chamber development observation along the well pair length and the cross sections in the mid, toe and heel areas. In addition, the cumulative recovery in given time, steam-to-oil ratio and CDOR （calendar day oil rate） production data are used to backup the observations produced in the simulated model. Furthermore, an additional component developed during the research is a statistical modification of a layer cake model with which to create a heterogeneous reservoir to represent real reservoir conditions, based on Monte Carlo＇s simulation.

Digital Twin-Assisted Knowledge Distillation Framework for Heterogeneous Federated Learning

In this paper,to deal with the heterogeneity in federated learning(FL)systems,a knowledge distillation(KD)driven training framework for FL is proposed,where each user can select its neural network model on demand and distill knowledge from a big teacher model using its own private dataset.To overcome the challenge of train the big teacher model in resource limited user devices,the digital twin(DT)is exploit in the way that the teacher model can be trained at DT located in the server with enough computing resources.Then,during model distillation,each user can update the parameters of its model at either the physical entity or the digital agent.The joint problem of model selection and training offloading and resource allocation for users is formulated as a mixed integer programming(MIP)problem.To solve the problem,Q-learning and optimization are jointly used,where Q-learning selects models for users and determines whether to train locally or on the server,and optimization is used to allocate resources for users based on the output of Q-learning.Simulation results show the proposed DT-assisted KD framework and joint optimization method can significantly improve the average accuracy of users while reducing the total delay.

Generation of optimal persistent formations for heterogeneous multi-agent systems with a leader constraint

In this study, we consider the generation of optimal persistent formations for heterogeneous multi-agent systems, with the leader constraint that only specific agents can act as leaders. We analyze three modes to control the optimal persistent formations in two-dimensional space, thereby establishing a model for their constrained generation. Then, we propose an algorithm for generating the optimal persistent formation for heterogeneous multi-agent systems with a leader constraint （LC-HMAS-OPFGA）, which is the exact solution algorithm of the model, and we theoretically prove its validity. This algorithm includes two kernel sub-algorithms, which are optimal persistent graph generating algorithm based on a minimum cost arborescence and the shortest path （MCA-SP-OPGGA）, and the optimal persistent graph adjusting algorithm based on the shortest path （SP-OPGAA）. Under a given agent formation shape and leader constraint, LC-HMAS-OPFGA first generates the network topology and its optimal rigid graph corresponding to this formation shape. Then, LC-HMAS- OPFGA uses MCA-SP-OPGGA to direct the optimal rigid graph to generate the optimal persistent graph. Finally, LC- HMAS-OPFGA uses SP-OPGAA to adjust the optimal persistent graph until it satisfies the leader constraint. We also demonstrate the algorithm, LC-HMAS-OPFGA, with an example and verify its effectiveness.

Efficiency-first spraying mission arrangement optimization with multiple UAVs in heterogeneous farmland with varying pesticide requirements

Combining multiple crop protection Unmanned Aerial Vehicles(UAVs)as a team for a scheduled spraying mission over farmland now is a common way to significantly increase efficiency.However,given some issues such as different configurations,irregular borders,and especially varying pesticide requirements,it is more important and more complex than other multi-Agent Systems(MASs)in common use.In this work,we focus on the mission arrangement of UAVs,which is the foundation of other high-level cooperations,systematically propose Efficiency-first Spraying Mission Arrangement Problem(ESMAP),and try to construct a united problem framework for the mission arrangement of crop protection UAVs.Besides,to characterise the differences in sub-areas,the varying pesticide requirement per unit is well considered based on Normalized Difference Vegetation Index(NDVI).Firstly,the mathematical model of multiple crop-protection UAVs is established and ESMAP is defined.Furthermore,an acquisition method of a farmland’s NDVI map is proposed,and the calculation method of pesticide volume based on NDVI is discussed.Secondly,an improved Genetic Algorithm(GA)is proposed to solve ESMAP,and a comparable combination algorithm is introduced.Numerical simulations for algorithm analysis are carried out within MATLAB,and it is determined that the proposed GA is more efficient and accurate than the latter.Finally,a mission arrangement tested with three UAVs was carried out to validate the effectiveness of the proposed GA in spraying operation.Test results illustrated that it performed well,which took only 90.6%of the operation time taken by the combination algorithm.

Distributed Heterogeneous Multi-Agent Optimization with Stochastic Sub-Gradient

This paper studies the optimization problem of heterogeneous networks under a timevarying topology.Each agent only accesses to one local objective function,which is nonsmooth.An improved algorithm with noisy measurement of local objective functions' sub-gradients and additive noises among information exchanging between each pair of agents is designed to minimize the sum of objective functions of all agents.To weaken the effect of these noises,two step sizes are introduced in the control protocol.By graph theory,stochastic analysis and martingale convergence theory,it is proved that if the sub-gradients are uniformly bounded,the sequence of digraphs is balanced and the union graph of all digraphs is joint strongly connected,then the designed control protocol can force all agents to find the global optimal point almost surely.At last,the authors give some numerical examples to verify the effectiveness of the stochastic sub-gradient algorithms.

Distributed Optimization for Heterogenous Second⁃Order Multi⁃Agent Systems

A continuous⁃time distributed optimization was researched for second⁃order heterogeneous multi⁃agent systems.The aim of this study is to keep the velocities of all agents the same and make the velocities converge to the optimal value to minimize the sum of local cost functions.First,an effective distributed controller which only uses local information was designed.Then,the stability and optimization of the systems were verified.Finally,a simulation case was used to illustrate the analytical results.

Optimal Control of Heterogeneous-Susceptible-Exposed-Infectious-Recovered-Susceptible Malware Propagation Model in Heterogeneous Degree-Based Wireless Sensor Networks

Heterogeneous wireless sensor networks(HWSNs)are vulnerable to malware propagation,because of their low configuration and weak defense mechanism.Therefore,an optimality system for HWSNs is developed to suppress malware propagation in this paper.Firstly,a heterogeneous-susceptible-exposed-infectious-recovered-susceptible(HSEIRS)model is proposed to describe the state dynamics of heterogeneous sensor nodes(HSNs)in HWSNs.Secondly,the existence of an optimal control problem with installing antivirus on HSNs to minimize the sum of the cumulative infection probabilities of HWSNs at a low cost based on the HSEIRS model is proved,and then an optimal control strategy for the problem is derived by the optimal control theory.Thirdly,the optimal control strategy based on the HSEIRS model is transformed into corresponding Hamiltonian by the Pontryagin’s minimum principle,and the corresponding optimality system is derived.Finally,the effectiveness of the optimality system is validated by the experimental simulations,and the results show that the infectious HSNs will fall to an extremely low level at a low cost.

Improved Performance for a Heterogeneous Satellite-Cooperative Network with Best Relay Node Selection

Nowadays, hybrid satellite-terrestrial cooperative network has emerged as a key technology to provide a great variety of communication services. The deployment of this network will improve coverage and capacity in remote areas. Despite the benefits of this network, by increasing the number of users, communication efficiency based on interference management is a major challenge in satellite-based system. Also, the direct links between satellite system and the terrestrial equipment do not always have desirable channel condition. In order to avoid serious throughput degradation, choosing a cooperative relay node is very important. In this paper, Stackelberg game is exploited for interference management that is raised by satellites in down link over terrestrial equipment. Then, for interference management between ground station and relay node with other mobile users, CVX is used to allocate optimum power. Also, the best relay node in this structure is selected based on the harmonic mean function. Thus, the performance of the heterogeneous satellite-cooperative network is investigated based on three benchmarks, namely, successful transmission, energy consumption and outage probability. Finally, the simulation results showed the effect of proposed system model on the performance of next generation satellite networks.

Joint Resource Allocation Using Evolutionary Algorithms in Heterogeneous Mobile Cloud Computing Networks

The problem of joint radio and cloud resources allocation is studied for heterogeneous mobile cloud computing networks. The objective of the proposed joint resource allocation schemes is to maximize the total utility of users as well as satisfy the required quality of service(QoS) such as the end-to-end response latency experienced by each user. We formulate the problem of joint resource allocation as a combinatorial optimization problem. Three evolutionary approaches are considered to solve the problem: genetic algorithm(GA), ant colony optimization with genetic algorithm(ACO-GA), and quantum genetic algorithm(QGA). To decrease the time complexity, we propose a mapping process between the resource allocation matrix and the chromosome of GA, ACO-GA, and QGA, search the available radio and cloud resource pairs based on the resource availability matrixes for ACOGA, and encode the difference value between the allocated resources and the minimum resource requirement for QGA. Extensive simulation results show that our proposed methods greatly outperform the existing algorithms in terms of running time, the accuracy of final results, the total utility, resource utilization and the end-to-end response latency guaranteeing.

Understanding visual-auditory correlation from heterogeneous features for cross-media retrieval

Cross-media retrieval is an interesting research topic,which seeks to remove the barriers among different modalities.To enable cross-media retrieval,it is needed to find the correlation measures between heterogeneous low-level features and to judge the semantic similarity.This paper presents a novel approach to learn cross-media correlation between visual features and auditory features for image-audio retrieval.A semi-supervised correlation preserving mapping(SSCPM)method is described to construct the isomorphic SSCPM subspace where canonical correlations between the original visual and auditory features are further preserved.Subspace optimization algorithm is proposed to improve the local image cluster and audio cluster quality in an interactive way.A unique relevance feedback strategy is developed to update the knowledge of cross-media correlation by learning from user behaviors,so retrieval performance is enhanced in a progressive manner.Experimental results show that the performance of our approach is effective.

Vertical handoff algorithm in heterogeneous networks to maximize system benefit

A new vertical handoff decision algorithm is proposed to maximize the system benefit in heterogeneous wireless networks which comprise cellular networks and wireless local area networks （WLANs）. Firstly the block probability, the drop probability and the number of users in the heterogeneous networks are calculated in the channel-guard call admission method, and a function of the system benefit which is based on the new call arrival rate and the handoff call arrival rate is proposed. Then the optimal radius of WLAN is obtained by using simulation annealing （SA） method to maximize the benefit. All the nodes should handoff from cellular network to WLAN if they enter WLAN＇s scope and handoff from WLAN to cellular network if they leave the scope. Finally, the algorithm in different new call arrival rates and handoff call arrival rates is analyzed and results show that it can achieve good effects.

PARALLEL IMPLEMENTATION AND OPTIMIZATION OF THE SEBVHOS ALGORITHM

In this paper, a parallel Surface Extraction from Binary Volumes with Higher-Order Smoothness (SEBVHOS) algorithm is proposed to accelerate the SEBVHOS execution. The original SEBVHOS algorithm is parallelized first, and then several performance optimization techniques which are loop optimization, cache optimization, false sharing optimization, synchronization overhead op-timization, and thread affinity optimization, are used to improve the implementation's performance on multi-core systems. The performance of the parallel SEBVHOS algorithm is analyzed on a dual-core system. The experimental results show that the parallel SEBVHOS algorithm achieves an average of 1.86x speedup. More importantly, our method does not come with additional aliasing artifacts, com-paring to the original SEBVHOS algorithm.

Context-Oriented Multi-RAT User Association and Resource Allocation with Triple Decision in 5G Heterogeneous Networks

In the upcoming 5 G heterogeneous networks, leveraging multiple radio access technologies(RATs) shows to be a crucial issue in achieving RAT multiplexing gain to meet the explosive traffic demand. For always best connection(ABC), users tend to activate parallel transmission across all available RATs. However from a system-wide perspective, this might not be optimal given the context of network load, interference and diverse service requirements. To intelligently determine how to use these multi-RAT access resources concurrently, this paper proposes a joint multi-RAT user association and resource allocation strategy with triple decision and integrated context awareness of users and networks. A dynamic game based ant colony algorithm(GACA) is designed to simultaneously maximize the system utility and the fairness of resource allocation. Simulation results show that it's more reasonable to make multi-RAT association decision from a system-wide viewpoint than from an individual one. Compared to max-SNR based and ABC based strategies, the proposed method alleviates network congestion and optimizes resource allocation. It obtains 39%~70% performance improvement.