非最小相位逐维定位多模型自适应解耦控制器

针对多变量系统中多个参数同时跳变导致模型数目巨大,计算时间长等问题,提出了采用逐维定位技术的多模型自适应解耦控制器.该方法将多维空间的并行寻优问题转化为多个一维空间的串行寻优问题,可大大减少系统模型集的数量.针对非最小相位系统,将系统的耦合作用视为可测干扰,采用前馈方法予以消除.最后给出全局收敛性分析.仿真结果表明当采用相同数目的模型时,其控制效果明显优于常规多模型控制器.

     

采用逐维定位的多模型自适应解耦控制器

针对多变量系统中多个参数同时变化导致模型数目巨大,计算时间长等问题,提出了采用逐维定位的多模型自适应解耦控制器．该方法将多维空间的并行寻优问题转化为多个一维空间的串行寻优问题,每一次固定其他参数、只针对一个参数寻找最优模型,可大大减少系统模型集的数量．该控制器基于性能指标搜索最优模型,通过加权多项式矩阵的选择,不但消除了稳态误差,任意配置闭环系统的极点,而且实现了动态解耦控制．最后给出全局收敛性分析．仿真结果表明当采用相同的固定模型覆盖每个参数的变化区间时,其模型集的数目远远小于常规多模型控制器．而当采用相同数目的模型时,其控制效果明显优于常规多模型控制器．     

一类非最小相位系统分层递阶多模型解耦控制器

针对多模型控制方法中模型数目巨大,计算时间长等问题,提出了分层递阶结构多模型自适应前馈解耦控制器.该控制器中固定模型集采用分层递阶结构,基于切换准则逐层搜索最优模型,动态构造下一层固定模型集实现完全覆盖.最后一层添加自适应模型消除稳态误差.针对非最小相位系统,将系统的耦合作用视为可测干扰,采用前馈方法予以消除.最后给出全局收敛性分析.仿真结果表明,与常规多模型控制方法相比,极大地减少了固定模型的数量.而当模型数目相同时,系统的暂态响应、解耦效果得到极大改善.     

多模型自适应控制的分层递阶构造与覆盖性质分析

针对一类非最小相位系统,设计一种多模型自适应控制器.该控制器由固定控制器模型、常规自适应模型和可重新赋值自适应模型构成.固定控制器模型采用分层递阶结构用来减少模型集的数目,根据切换指标选出的上一层最优控制器,动态设计本层固定控制器模型实现对其参数变化范围的覆盖.该控制器采用直接自适应算法,通过加权多项式的选取,消除了稳态误差.文末对系统的覆盖性、模型数目等进行了分析.仿真结果表明当采用相同数目的模型时,其控制效果明显优于常规多模型控制器.     

基于切换的多模型二阶段自适应控制器设计

针对一类具有参数跳变特性的离散时间系统,设计一类基于切换策略的新型多模型二阶段自适应控制器.该控制器首先将系统不确定参数的变化空间划分为多个子空间,在每个空间内建立多个自适应模型.为了克服多模型退化,保持模型的多样性以应对参数跳变,采用带约束的二阶段自适应方法对未知参数进行实时估计,并据此设计相应的子控制器;然后基于切换策略,选取该时刻的最优子控制器作为系统的控制器,从而减小系统暂态误差,提高系统动态性能;最后进行数值仿真研究,仿真结果表明该控制器结合了切换机制和二阶段自适应的优点,在相同模型数量的情况下,能够快速逼近参数跳变以后系统新的工作点,显著地缩短系统的过渡过程,提高暂态性能.     

基于多层蚁群算法的变论域模糊控制

针对模糊控制器控制精度不高、自适应能力有限等问题,提出一种变论域自适应模糊控制方式.首先在对离散蚁群算法改进的基础上,提出用于连续域寻优的多层蚁群算法.其通过将解空间分成有限网格,并且算法在迭代过程中采用三个阶段的搜索策略,每个阶段采用异构搜索机制.然后根据系统性能利用改进算法动态调整伸缩因子,从而构成基于多层蚁群算法的变论域自适应模糊控制器.最后将此控制器用于中厚板液压位置伺服系统中.仿真结果表明,采用自适应模糊控制器的伺服系统收敛速度明显加快,此控制策略在适应能力与鲁棒性好于其它控制方式.     

近空间飞行器的多模型切换控制

针对受到扰动的变机翼后掠角近空间飞行器,研究一类基于多模型切换的多输入多输出非线性系统的模糊自适应鲁棒控制器的设计问题.通过构造公共Lyapunov函数设计系统的控制器,采用动态面控制方法避免了控制器设计中的计算膨胀问题,利用自适应模糊系统和鲁棒控制项在线消除系统中的未知干扰影响.仿真结果表明了该方法的有效性.     

多模型分层递阶自适应前馈解耦控制器

针对参数跳变系统,提出一种基于分层递阶结构的多模型自适应前馈解耦控制器.该控制器采用多模型方法来提高系统的暂态性能;采用自适应方法消除系统的稳态误差,采用分层递阶结构减少系统模型集的数量和计算时间.为了在分布式计算机集散控制系统(DCS)中得到应用,该控制器根据耦合的形成机理和DCS的结构特点,将系统变量之间的耦合作用视为可测干扰,采用前馈结构予以消除.通过加权多项式的选取,不仅实现了极点配置,而且可以动态解耦.最后给出了全局收敛性分析.仿真结果表明,与常规多模型控制方法相比,大大减少了固定模型的数量;而当模型数目相同时,系统的暂态响应、解耦效果都大为改善.     

基于有界扰动分区的多模型自适应控制

对于含有界扰动的离散时间系统,将有界扰动变化范围分割成若干小区间,针对这些区间建立多个自适应模型,并针对每一个自适应模型建立相应的控制器.给定一个指标切换函数,基于多个自适应模型控制器和给定的指标切换函数构成多模型自适应控制器.可以证明多模型自适应控制器能够保证闭环系统是输入输出有界稳定的.仿真实例表明,当被控对象的有界扰动随时间变化的时候,采用单一模型自适应控制,系统输出很难跟踪设定值,而多模型自适应控制器却极大地改善了控制品质.     

具有大不确定性对象的分层切换控制方法

针对大模型不确定性对象的控制问题,提出了一种基于鲁棒控制理论的多模型分层切换控制方法.为减少覆盖不确定性需要的模型数量,采用多个乘性不确定模型来描述对象,并应用LMI方法设计控制器集合.考虑鲁棒控制中常用系统增益来度量不确定性,设计了一种基于不确定性增益估计的切换指标函数,并据此将控制器集合中合适的控制器连接到反馈回路中.理论分析表明系统BIBO稳定,且具有一定的扰动抑制能力.仿真实验结果验证了控制方法的有效性.     

Modelling and optimization of grinding processes

The paper describes different methods for modelling and optimization of grinding processes. First the process and product quality characterizing quantities have to be measured. Afterwards different model types, e.g. physical–empirical basic grinding models as well as empirical process models based on neural networks, fuzzy set theory and standard multiple regression methods, are discussed for an off-line process conceptualization and optimization using a genetic algorithm. The assessment of grinding process results, which build the individuals in the genetic algorithm's population, is carried out using a target tree method. The methods presented are integrated into an existing grinding information system, which is part of a three control loop system for quality assurance.     

一类具有羊群效应的多重多维泰勒网动力学模型

含大量有主观判断力个体参与的系统,往往会表现出羊群效应的特点,现有研究方法主要从机理和个体角度展开,不适合建立整体模型.本文提出带间歇反馈的多重多维泰勒网动力学模型建模方法,适合建立含有羊群效应系统的整体优化模型.首先讨论了多维泰勒网和间歇反馈模型建立方法.然后研究了多重多维泰勒网的特性,并通过多重多维泰勒网调节羊群效应和系统长期趋势两个部分在系统中合理比重.最后给出了辨识模型参数的具体方法和步骤.应用实例的结果验证了带间歇反馈的多重多维泰勒网建模方法在实际应用中的可实现性,同时其拥有更好的预测精度.     

Personalized Hand Modeling from Multiple Postures with Multi-View Color Images

Personalized hand models can be utilized to synthesize high quality hand datasets, provide more possible training data for deep learning and improve the accuracy of hand pose estimation. In recent years, parameterized hand models, e.g., MANO, are widely used for obtaining personalized hand models. However, due to the low resolution of existing parameterized hand models, it is still hard to obtain high-fidelity personalized hand models. In this paper, we propose a new method to estimate personalized hand models from multiple hand postures with multi-view color images. The personalized hand model is represented by a personalized neutral hand, and multiple hand postures. We propose a novel optimization strategy to estimate the neutral hand from multiple hand postures. To demonstrate the performance of our method, we have built a multi-view system and captured more than 35 people, and each of them has 30 hand postures. We hope the estimated hand models can boost the research of high-fidelity parameterized hand modeling in the future. All the hand models are publicly available on www.yangangwang.com .     

改进MOPSO在物流节点选址模型中的应用

为了降低物流系统的运营成本,提高物流系统的运作效率,构建了物流系统运营成本最小以及顾客时间满意度最大的多目标物流节点选址模型,并在模型求解过程中针对多目标粒子群算法的不足,从外部存档的更新、粒子学习样本的选择以及粒子的变异三个方面进行改进,将改进的多目标粒子群算法用于物流节点选址模型的求解。仿真结果表明,改进的算法相较于其他优化算法,具有较好的分布性和收敛性。     

动态优化双估计器的多模型自适应混合控制

针对参数子集个数较多导致计算量较大和由于系统参数发生跳变造成系统暂态性能差的问题,本文提出了基于动态优化双估计器的多模型自适应混合控制方法.首先对多个参数子集进行动态优化得到最优参数子集,减少了需要计算的模型数量,提高了系统收敛速度;其次对被控对象设置一个固定初值的估计器和一个可重新赋值的估计器,固定估计器用于初始时刻对参数的估计,可赋值估计器动态调整估计初值用于减小估计误差,提高系统暂态性能.最后的仿真结果表明了该方法的有效性,并给出了系统的稳定性及收敛性分析.     

Efficient design optimization strategy for structural dynamic systems using a reduced basis method combined with an equivalent static load

This paper presents an efficient structural design optimization strategy that combines the reduced basis method (RBM) with the equivalent static load (ESL). In dynamic response optimization using ESL, the computation of a static system optimization is repeatedly executed under multiple static loads. In this process, we propose parametrizing the static system and employing the RBM with global proper orthogonal decomposition (POD). In general, the snapshots for the sampling procedure under multiple loads increase proportionally to the number of loads, which results in an inefficient computational procedure. Thus, we propose taking snapshots with a proper orthogonal mode (POM) of the multiple loads rather than for the multiple loads themselves. The number of snapshots then decreases, and the original system can be efficiently reduced. We directly employ the framework of the proposed RBM with the POM of multiple loads to ESL-based design optimization, and the results indicate that the proposed method is more efficient than conventional ESL-based optimization as well as a full order model. Various numerical examples, including comparisons of relative errors and the dynamic response optimizations, support the strength of the proposed strategy.     

Toward Global Optimization of Case-Based Reasoning Systems for Financial Forecasting

This paper presents a simultaneous optimization method of a case-based reasoning (CBR) system using a genetic algorithm (GA) for financial forecasting. Prior research proposed many hybrid models of CBR and the GA for selecting a relevant feature subset or optimizing feature weights. Most research used the GA for improving only a part of architectural factors of the CBR model. However, the performance of the CBR model may be enhanced when these factors are simultaneously considered. In this study, the GA simultaneously optimizes multiple factors of the CBR system. Experimental results show that a GA approach to simultaneous optimization of the CBR model outperforms other conventional approaches for financial forecasting.     

Cross-layer optimization for wireless mesh networks with smart antennas

A cross-layer optimization framework for wireless mesh networks is presented where at each node, various smart antenna techniques such as beam-forming, spatial division multiple access and spatial division multiplexing are employed. These techniques provide interference suppression, capability for simultaneous communication with several nodes and transmission with higher data rates, respectively, through multiple antennas. By integrating different combinations of the multi-antenna techniques in physical layer with various constraints from MAC and network layers, three Mixed Integer Linear Programming (MILP) models are presented to minimize the system activation time. Since these optimization problems are complex combinatorial, the optimal solution is approached by a Column Generation decomposition method. The numerical results for different network scenarios with various node densities, number of antennas, transmission ranges and number of sessions are provided. It is shown that the resulted directive, multiple access and multiplexing gains combined with scheduling, effectively increase both the spectrum spatial reuse and the capacity of the links and therefore, enhance the achievable system throughput. Our cross-layer approach is also extended to consider heterogeneous networks and we present a multi-criteria optimization framework to model the design problem where the objective is to jointly minimize the cost of deployment and the system activation time. Our results reveal the benefits of joint design in terms of reducing the cost of deployment while achieving higher system performance.     

基于划分的蚁群算法求解货物权重车辆路径问题

考虑单产品分销网络中的车辆路径问题(VRP:vehicle routing problem).与以往诸多研究不同的是,建立了一种带货物载重量的VRP模型(weighted VRP),即车辆在两个顾客之间行驶时的载重量也作为影响运输费用的一个因素考虑.因此,需求量较大的顾客拥有较高的车辆运输优先权.在分析了问题性质的基础上,提出一种基于划分策略的蚁群算法PMMAS求解货物权重车辆路径问题,并与其他常用的启发式算法进行比较分析,表明了算法的有效性.