Edge-Federated Self-Supervised Communication Optimization Framework Based on Sparsification and Quantization Compression

The federated self-supervised framework is a distributed machine learning method that combines federated learning and self-supervised learning, which can effectively solve the problem of traditional federated learning being difficult to process large-scale unlabeled data. The existing federated self-supervision framework has problems with low communication efficiency and high communication delay between clients and central servers. Therefore, we added edge servers to the federated self-supervision framework to reduce the pressure on the central server caused by frequent communication between both ends. A communication compression scheme using gradient quantization and sparsification was proposed to optimize the communication of the entire framework, and the algorithm of the sparse communication compression module was improved. Experiments have proved that the learning rate changes of the improved sparse communication compression module are smoother and more stable. Our communication compression scheme effectively reduced the overall communication overhead.

Dilemmas and Optimization of Third-Party Assessment System in Chinese Fair Competition Review

The current predominant self-review mechanism by policy-making bodies suffers from deficiencies such as insufficient motivations, limited review capabilities, and weak external supervision. Third-party assessment, characterized by independence and specialization, is designed to mitigate these shortcomings. However, the implementation of third-party assessment faces challenges too. This paper intends to improve the third-party assessment system and to realize the legislative purpose of the system. Based on social research, discussions and exchanges with relevant parties, and the existing research results, this paper analyzes the challenges and possible optimization measures for the third-party assessment. The challenges include repulsion from policy-making bodies, insufficient independence of assessment bodies, disparity of assessment quality, and limited application of assessment outcomes. Possible optimization measures include promoting fair competition culture, increasing the acceptance of third-party assessment from policy-making bodies, enhancing the quality of third-party assessment, clarifying the relationship between policy-making bodies and assessment bodies, ensuring the independence of third-party assessments, and promoting the application of assessment results.

Real-valued multi-area self set optimization in immunity-based network intrusion detection system

The real-valued self set in immunity-based network intrusion detection system （INIDS） has some defects： multi-area and overlapping, which are ignored before. The detectors generated by this kind of self set may have the problem of boundary holes between self and nonself regions, and the generation efficiency is low, so that, the self set needs to be optimized before generation stage. This paper proposes a self set optimization algorithm which uses the modified clustering algorithm and Gaussian distribution theory. The clustering deals with multi-area and the Gaussian distribution deals with the overlapping. The algorithm was tested by Iris data and real network data, and the results show that the optimized self set can solve the problem of boundary holes, increase the efficiency of detector generation effectively, and improve the system＇s detection rate.

Optimization of Planar Five-bar Parallel Mechanism via Self-reconfiguration Method

The parallel mechanisms have the disadvantage of small workspace and complication in kinematics and dynamics. An optimizing design for the parallel mechanisms can improve the motion performance relatively, but not guarantee the design results which satisfy the various practical requirements simultaneously. In this paper, a dynamical and optimal synthesis method is proposed for parallel mechanisms based on the dynamical reconfiguration technique. As a specific, application, the problem of optimizing the kinematics isotropy of a five-bar planar parallel mechanism is studied. The motion of a reconfigurable mechanism can be parted into two phases, the natural motion phase and the reconfiguration phase. The two motion phases can be studied by the same performance evaluation methodology. This points out from both theory and practices a novel method for improving the motion performance of the parallel mechanisms. Simulation by a symmetrical five-bar planar parallel manipulator shows some aspects of the investigations.

Prediction-based Manufacturing Center Self-adaptive Demand Side Energy Optimization in Cyber Physical Systems

Cyber physical systems（CPS） recently emerge as a new technology which can provide promising approaches to demand side management（DSM）, an important capability in industrial power systems. Meanwhile, the manufacturing center is a typical industrial power subsystem with dozens of high energy consumption devices which have complex physical dynamics. DSM, integrated with CPS, is an effective methodology for solving energy optimization problems in manufacturing center. This paper presents a prediction-based manufacturing center self-adaptive energy optimization method for demand side management in cyber physical systems. To gain prior knowledge of DSM operating results, a sparse Bayesian learning based componential forecasting method is introduced to predict 24-hour electric load levels for specific industrial areas in China. From this data, a pricing strategy is designed based on short-term load forecasting results. To minimize total energy costs while guaranteeing manufacturing center service quality, an adaptive demand side energy optimization algorithm is presented. The proposed scheme is tested in a machining center energy optimization experiment. An AMI sensing system is then used to measure the demand side energy consumption of the manufacturing center. Based on the data collected from the sensing system, the load prediction-based energy optimization scheme is implemented. By employing both the PSO and the CPSO method, the problem of DSM in the manufac^ring center is solved. The results of the experiment show the self-adaptive CPSO energy optimization method enhances optimization by 5% compared with the traditional PSO optimization method.

Enhanced self-adaptive evolutionary algorithm for numerical optimization

There are many population-based stochastic search algorithms for solving optimization problems. However, the universality and robustness of these algorithms are still unsatisfactory. This paper proposes an enhanced self-adaptiveevolutionary algorithm （ESEA） to overcome the demerits above. In the ESEA, four evolutionary operators are designed to enhance the evolutionary structure. Besides, the ESEA employs four effective search strategies under the framework of the self-adaptive learning. Four groups of the experiments are done to find out the most suitable parameter values for the ESEA. In order to verify the performance of the proposed algorithm, 26 state-of-the-art test functions are solved by the ESEA and its competitors. The experimental results demonstrate that the universality and robustness of the ESEA out-perform its competitors.

Particle Swarm Optimization Algorithm Based on Chaotic Sequences and Dynamic Self-Adaptive Strategy

To deal with the problems of premature convergence and tending to jump into the local optimum in the traditional particle swarm optimization, a novel improved particle swarm optimization algorithm was proposed. The self-adaptive inertia weight factor was used to accelerate the converging speed, and chaotic sequences were used to tune the acceleration coefficients for the balance between exploration and exploitation. The performance of the proposed algorithm was tested on four classical multi-objective optimization functions by comparing with the non-dominated sorting genetic algorithm and multi-objective particle swarm optimization algorithm. The results verified the effectiveness of the algorithm, which improved the premature convergence problem with faster convergence rate and strong ability to jump out of local optimum.

基于CNN-SAEDN-Res的短期电力负荷预测方法

基于深度学习的序列模型难以处理混有非时序因素的负荷数据,这导致预测精度不足。提出一种基于卷积神经网络(CNN)、自注意力编码解码网络(SAEDN)和残差优化(Res)的短期电力负荷预测方法。特征提取模块由二维卷积神经网络组成,用于挖掘数据间的局部相关性,获取高维特征。初始负荷预测模块由自注意力编码解码网络和前馈神经网络构成,利用自注意力机制对高维特征进行自注意力编码,获取数据间的全局相关性,从而模型能根据数据间的耦合关系保留混有非时序因素数据中的重要信息,通过解码模块进行自注意力解码,并利用前馈神经网络回归初始负荷。引入残差机制构建负荷优化模块,生成负荷残差,优化初始负荷。算例结果表明,所提方法在预测精度和预测稳定性方面具有优势。

一种POA-VMD和自编码器结合的风电机组轴承劣化指标构建及故障诊断方法

针对目前轴承性能劣化指标的构建及故障诊断高度依赖专家经验,限制条件繁多,实际应用情景单一的问题,提出一种鹈鹕优化算法(POA)优化的变分模态分解(VMD)和自编码器结合的风机轴承劣化指标构建及故障诊断方法。首先利用POA-VMD算法将轴承振动信号采用自适应方法分解为K个固有模态分量(IMF),并针对上述分量分别构建K个自编码器;然后以正常状态振动信号的分解结果为训练样本完成自编码器的训练,并以训练完成后模型的输出结果为基础构建轴承劣化指标,借助劣化指标监测轴承早期微弱故障;最后对故障时刻振动信号的IMF分量重构结果进行包络谱分析,确定故障的类型。经实验验证:该方法不仅可以清晰地展现轴承的劣化过程,对早期微弱故障敏感性高,而且在故障发生后可以准确诊断出故障类型。

Hydraulic Self Servo Swing Cylinder Structure Optimization and Dynamic Characteristics Analysis Based on Genetic Algorithm

The dynamic characteristics of hydraulic self servo swing cylinder were analyzed according to the hydraulic system natural frequency formula. Based on that,a method of the hydraulic self servo swing cylinder structure optimization based on genetic algorithm was proposed in this paper. By analyzing the four parameters that affect the dynamic characteristics, we had to optimize the structure to obtain as larger the Dm( displacement) as possible under the condition with the purpose of improving the dynamic characteristics of hydraulic self servo swing cylinder. So three state equations were established in this paper. The paper analyzed the effect of the four parameters in hydraulic self servo swing cylinder natural frequency equation and used the genetic algorithm to obtain the optimal solution of structure parameters. The model was simulated by substituting the parameters and initial value to the simulink model. Simulation results show that: using self servo hydraulic swing cylinder natural frequency equation to study its dynamic response characteristics is very effective.Compared with no optimization,the overall system dynamic response speed is significantly improved.

SELF-RECONFIGURATION OF UNDERACTUATED REDUNDANT MANIPULATORS WITH OPTIMIZING THE FLEXIBILITY ELLIPSOID

The multi-modes feature, the measure of the manipulating flexibility, andself-reconfiguration control method of the underactuated redundant manipulators are investigatedbased on the optimizing technology. The relationship between the configuration of the joint spaceand the manipulating flexibility of the underactuated redundant manipulator is analyzed, a newmeasure of manipulating flexibility ellipsoid for the underactuated redundant manipulator withpassive joints in locked mode is proposed, which can be used to get the optimal configuration forthe realization of the self-reconfiguration control. Furthermore, a time-varying nonlinear controlmethod based on harmonic inputs is suggested for fulfilling the self-reconfiguration. A simulationexample of a three-DOFs underactuated manipulator with one passive joint features some aspects ofthe investigations.

Topology optimization of 3D structures with design-dependent loads

Topology optimization of continuum structures with design-dependent loads has long been a challenge. In this paper, the topology optimization of 3D structures subjected to design-dependent loads is investigated. A boundary search scheme is proposed for 3D problems, by means of which the load surface can be identified effectively and efficiently, and the difficulties arising in other approaches can be overcome. The load surfaces are made up of the boundaries of finite elements and the loads can be directly applied to corresponding element nodes, which leads to great convenience in the application of this method. Finally, the effectiveness and efficiency of the proposed method is validated by several numerical examples.

OPTIMIZING DESIGN OF MECHANICAL SELF-CENTERING DEVICE FOR SUSPENSION HEIGHT

Firstly, in view of the respective defects of existing self-centering devices for vehicle suspension height, the design scheme of the proposed mechanical self-centering device for suspension height is described. Taking the rear suspension of a certain light bus as a research example, the structures and parameters of the novel device are designed and ascertained. Then, the road excitation models, the performance evaluation indexes and the half-vehicle model are built, the simulation outputs of time and frequency domain are obtained with the road excitations of random and pulse by using MATLAB/Simulink software. So the main characteristics of the self-centering suspension are presented preliminarily. Finally, a multi-objective parameter design optimization model for the self-centering device is built by weighted sum approach, and optimal solution is obtained by adopting complex approach. The relevant choosing-type parameters for self-centering device components are deduced by using discrete variable optimal method, and the optimal results are verified and analyzed. So the performance potentials of the self-centering device are exerted fully in condition of ensuring overall suspension performances.

基于改进哈里斯鹰和B-spline曲线的无人机路径规划研究

针对无人机在动态环境中的全局路径规划问题,提出了一种改进哈里斯鹰优化算法。针对算法后期搜索性能不足等问题,提出自适应混沌和核心种群动态划分策略,提高算法后期的搜索能力;修改哈里斯鹰更新公式,引入黄金正弦策略,提高算法搜索效率;融合自适应动态云最优解扰动策略,提高算法跳出局部极值的能力;针对三维栅格路径规划问题,设置了一种估值函数,通过计算栅格到达终点的代价,帮助算法进行节点筛选,使算法能搜索到更短路径,并针对路径转角不平滑的问题,使用3次B-spline曲线对路径转角进行处理,使路径更适合无人机飞行。通过国际标准测试函数和在不同大小、不同复杂程度的静态、动态栅格地图进行仿真实验。实验结果显示,本文算法相较于对比算法,规划出的路径平均缩短了14.94%、转角数量平均减少了53.31%。

Dynamic Control and Optimization of Capital-Constrained Stochastic Inventory Systems

For most firms,especially the small-and medium-sized ones,the operational decisions are affected by their internal capital and ability to obtain external capital.However,the majority of the current studies on dynamic inventory control ignore the firm’s financial status and financing issues completely.An important question that arises is:what are the dynamic optimal inventory and financing policies for firms with limited capital and limited access to external capital?In this paper,we review some of the latest developments in this area.After a brief review of single period models,we focus on multi-period dynamic control of the firm who aims to optimize its xpected terminal wealth.Two cases are discussed in detail:self-finance and short term finance.In the first case,the firm has to rely on its own capital for all ordering decisions,while in the second,the firm can borrow short term loan from lenders.A detailed characterization of the optimal policy is presented and its managerial insights are discussed.Several possible extensions are suggested.

Self-adaptive strategy for one-dimensional finite element method based on EEP method with optimal super-convergence order

Based on the newly-developed element energy projection （EEP） method with optimal super-convergence order for computation of super-convergent results, an improved self-adaptive strategy for one-dimensional finite element method （FEM） is proposed. In the strategy, a posteriori errors are estimated by comparing FEM solutions to EEP super-convergent solutions with optimal order of super-convergence, meshes are refined by using the error-averaging method. Quasi-FEM solutions are used to replace the true FEM solutions in the adaptive process. This strategy has been found to be simple, clear, efficient and reliable. For most problems, only one adaptive step is needed to produce the required FEM solutions which pointwise satisfy the user specified error tolerances in the max-norm. Taking the elliptical ordinary differential equation of the second order as the model problem, this paper describes the fundamental idea, implementation strategy and computational algorithm and representative numerical examples are given to show the effectiveness and reliability of the proposed approach.

基于SOM-FCM和KELM组合方法的短期光伏功率预测

为了提高短期光伏发电预测的精度,本文提出了一种将聚类后的自组织映射网络(SOM)与优化的核极值学习机(KELM)方法相结合的混合预测模型。首先,利用SOM来对训练数据集进行初始划分。然后,利用模糊C均值(FCM)对训练好的SOM网络进行聚类操作,同时利用Davies-Bouldin指数(DBI)来确定最佳聚类的大小。最后,在每个数据分区中,通过结合差分演化算法优化的KELM方法来建立区域KELM模型,或者结合最小二乘估计的多元线性回归(MR)方法来构建区域MR模型。此外,本文还提出了基于SOM的不同局部多元回归模型。将提出的结合SOM-FCM和KELM的混合预测模型分别应用于GEFCom2014三个不同太阳能电站,进行提前一小时的发电功率预测。与其他预测模型相比,光伏发电站1的平均绝对误差(MAE)降低了61.41%,光伏发电站2的MAE降低了60.19%,光伏发电站3的MAE降低了58.92%。光伏发电站1的均方根误差(RMSE)降低了52.06%,光伏发电站2的RMSE降低了54.56%,光伏发电站3的RMSE降低了51.43%。实验结果表明,提出的结合SOMFCM和KELM的方法可显著提高预测准确性。

Improved particle swarm optimization algorithm for multi-reservoir system operation

In this paper, a hybrid improved particle swarm optimization （IPSO） algorithm is proposed for the optimization of hydroelectric power scheduling in multi-reservoir systems. The conventional particle swarm optimization （PSO） algorithm is improved in two ways： （1） The linearly decreasing inertia weight coefficient （LDIWC） is replaced by a self-adaptive exponential inertia weight coefficient （SEIWC）, which could make the PSO algorithm more balanceable and more effective in both global and local searches. （2） The crossover and mutation idea inspired by the genetic algorithm （GA） is imported into the particle updating method to enhance the diversity of populations. The potential ability of IPSO in nonlinear numerical function optimization was first tested with three classical benchmark functions. Then, a long-term multi-reservoir system operation model based on IPSO was designed and a case study was carried out in the Minjiang Basin in China, where there is a power system consisting of 26 hydroelectric power plants. The scheduling results of the IPSO algorithm were found to outperform PSO and to be comparable with the results of the dynamic programming successive approximation （DPSA） algorithm.

Al_2O_3-2SiO_2 Nanoparticles with Defined Al-Si Ratio:Processing Optimization and Conversion

Attempts had been made to synthesize Al2O3-2SiO2 nanopowders by sol-gel method with tetraethoxysilane（TEOS） and aluminum nitrate（ANN） as the starting materials.DTS,TEM,SEM and BET were employed to study the effects of process parameters on the size,specific surface area and structure（morphology） of powders.The alkali-activation reactivity of the powders was tested for manufacturing geopolymers and their hydrothermal reactions were performed for fabricating zeolites.The results show that the optimum process parameters and drying method for preparing Al2O3-2SiO2 nanopowders are as follows：the molar ratio of water and ethanol to TEOS are 0：1 and 12：1 respectively at synthetic temperature of 50 ℃ and the drying method is azeotropic distillation with microwave drying.The average particle diameters of the powders were about 70 nm and the largest BET specific surface area was up to 669 m^2·g^-1.The compressive strength of the geopolymer and the calcium exchange capacity（by CaCO3） of NaA zeolite prepared with the powders reached to 29 MPa and 366 m^2·g^-1 respectively.

MODS： A Novel Metaheuristic of Deterministic Swapping for the Multi-Objective Optimization of Combinatorials Problems

This paper states a new metaheuristic based on Deterministic Finite Automata （DFA） for the multi - objective optimization of combinatorial problems. First, a new DFA named Multi - Objective Deterministic Finite Automata （MDFA） is defined. MDFA allows the representation of the feasible solutions space of combinatorial problems. Second, it is defined and implemented a metaheuritic based on MDFA theory. It is named Metaheuristic of Deterministic Swapping （MODS）. MODS is a local search strategy that works using a MDFA. Due to this, MODS never take into account unfeasible solutions. Hence, it is not necessary to verify the problem constraints for a new solution found. Lastly, MODS is tested using well know instances of the Bi-Objective Traveling Salesman Problem （TSP） from TSPLIB. Its results were compared with eight Ant Colony inspired algorithms and two Genetic algorithms taken from the specialized literature. The comparison was made using metrics such as Spacing, Generational Distance, Inverse Generational Distance and No-Dominated Generation Vectors. In every case, the MODS results on the metrics were always better and in some of those cases, the superiority was 100%.