GPU-based model predictive control for continuous casting spray cooling control system using particle swarm optimization

Model predictive control (MPC) for spray cooling control system requires a repeated online solution of an optimization problem that includes partial differential equations (PDEs). To simulate the future temperature behavior of steel billets, 3D dynamic heat transfer model is used. The special solution domain of PDEs has led to large computation cost, which is the main challenge in the real-time practical application of spray cooling control system. Meanwhile, the heat transfer coefficients need to be identified using the measured surface temperature. This work presents a two-level parallel solution method implemented on a Graphics processing unit (GPU) for MPC of spray cooling control systems and a weighted least squares modified conjugate gradient method (WLS–MCG) for identification of heat transfer coefficients. Two-level parallel solution method consists of parallel-based heat transfer model and stream parallel particle swarm optimization (PSO). PSO is used to solve the optimization problem. WLS–MCG consists of the weighted least squares (WLS) and modified conjugate gradient method (MCG). The experimental results show that the two-level parallel solution method has good computational performance and achieves satisfactory control performance.     

Nonlinear model predictive control with relevance vector regression and particle swarm optimization

In this paper, a nonlinear model predictive control strategy which utilizes a probabilistic sparse kernel learning technique called relevance vector regression （RVR） and particle swarm optimization with controllable random exploration velocity （PSO-CREV） is applied to a catalytic continuous stirred tank reactor （CSTR） process. An accurate reliable nonlinear model is first identified by RVR with a radial basis function （RBF） kernel and then the optimization of control sequence is speeded up by PSO-CREV. Additional stochastic behavior in PSO-CREV is omitted for faster convergence of nonlinear optimization. An improved system performance is guaranteed by an accurate sparse predictive model and an efficient and fast optimization algorithm. To compare the performance, model predictive control （MPC） using a deterministic sparse kernel learning technique called Least squares support vector machines （LS-SVM） regression is done on a CSTR. Relevance vector regression shows improved tracking performance with very less computation time which is much essential for real time control.     

控制系统的辨识建模及微粒群优化设计

针对控制系统的传递函数建模与控制器的参数优化问题,提出了基于Prony和微粒群优化（PSO）算法的设计方案。首先在被控对象的输入端施加一个脉冲信号,然后对其输出信号进行Prony分析,得出该被控对象的传递函数,最后采用改进PSO算法进行控制器的参数优化设计。基于辨识的Prony算法可快速准确得出被控对象的传递函数;基于T-S模型模糊自适应的改进PSO算法（T-SPSO算法）依据种群当前最优性能指标和惯性权重自适应惯性权重取值,较好解决了PSO算法的早熟问题,可以更好地优化控制器参数。该方案实现了控制系统的精确建模与优化设计,仿真结果验证了所提方案的有效性。     

Chaos particle swarm optimization and T-S fuzzy modeling approaches to constrained predictive control

Predictive control of systems is very much related to the efficiency and cost of systems, as well as to the quality of systems outcomes. However, it is difficult to achieve optimal predictive control because most predictive controls for systems have characteristics of randomness, strong and complex constraints, large delay time, fuzziness, and nonlinearity. Conventional methods of solving constrained nonlinear optimization problems for predictive control are mainly based on quadratic programming, which is quite sensitive to initial values, easy to trap in local minimal points, and requires large computational effort. In recent years, T-S fuzzy modeling has been found to be an effective approach in performing predictive control. Intelligent optimization algorithms, such as chaos optimization algorithm (COA) and particle swarm optimization (PSO), have been shown to have faster convergence and higher iterative accuracy than those based on conventional optimization methods. In this paper, chaos particle swarm optimization (CPSO), which involves combining the strengths of COA and PSO, and T-S fuzzy modeling are proposed as approaches to perform constrained predictive control. Predictive control of temperature of continued hyperthermic celiac perfusion for medical treatment based on the proposed approaches was carried out. Simulation tests were conducted to evaluate the performance of temperature control based on T-S fuzzy modeling and CPSO. Test results indicate that the T-S fuzzy model based on CPSO outperforms models based on generalized predictive control, COA, and PSO.     

Intelligent identification and control using improved fuzzy particle swarm optimization

This paper presents a novel improved fuzzy particle swarm optimization (IFPSO) algorithm to the intelligent identification and control of a dynamic system. The proposed algorithm estimates optimally the parameters of system and controller by minimizing the mean of squared errors. The particle swarm optimization is enhanced intelligently by using a fuzzy inertia weight to rationally balance the global and local exploitation abilities. In the proposed IFPSO, every particle dynamically adjusts inertia weight according to particles best memories using a nonlinear fuzzy model. As a result, the IFPSO algorithm has a faster convergence speed and a higher accuracy. The performance of IFPSO algorithm is compared with advanced algorithms such as Real-Coded Genetic Algorithm (RCGA), Linearly Decreasing Inertia Weight PSO (LDWPSO) and Fuzzy PSO (FPSO) in terms of parameter accuracy and convergence speed. Simulation results demonstrate the effectiveness of the proposed algorithm.     

基于分解的预测型动态多目标粒子群优化算法

针对动态多目标问题求解,提出一种基于分解的预测型动态多目标粒子群优化算法.首先借助分解思想,将目标问题划分为多个不同的子问题,当问题动态变化时,选择对应于不同子问题的优化个体检测环境变化程度,以提高算法对不同动态问题的适应与响应能力;然后,设计一种群体预测策略,通过将目标空间中相同收敛方向上不同时刻的个体位置转换为时间序列,引入时间序列预测方法预测下一刻位置,从而提高预测种群的多样性和有效性,进而有效减少算法在问题变化后的收敛时间;最后,为避免问题发生变化后个体与子问题不匹配,设计一种再匹配策略,以提高预测策略的准确性.实验结果表明,在6个标准动态多目标测试问题上,与2个动态多目标优化算法进行比较,所提出算法在收敛性、分布性与稳定性上均具有显著优势.     

Robust optimization using multi-objective particle swarm optimization

This article proposes an algorithm to search for solutions which are robust against small perturbations in design variables. The proposed algorithm formulates robust optimization as a bi-objective optimization problem, and fi nds solutions by multi-objective particle swarm optimization (MOPSO). Experimental results have shown that MOPSO has a better performance at fi nding multiple robust solutions than a previous method using a multi-objective genetic algorithm.     

多智能体粒子群优化的SVR 模型预测控制

参数的优化选择对支持向量回归机的预测精度和泛化能力影响显著,鉴于此,提出一种多智能体粒子群算法(MAPSO)寻优其参数的方法,并建立MAPSO支持向量回归模型,用于非线性系统的模型预测控制,推导出最优控制率.采用该算法对非线性系统进行仿真,并与基于粒子群算法、基于遗传算法优化支持向量回归机的模型预测控制方法和RBF神经网络的预测控制方法进行比较,结果表明,所提出的算法具有更好的控制性能,可以有效应用于非线性系统控制中.     

采用dsDNA-MC优化的非线性系统预测控制

将基于DNA双链结构的膜计算优化方法（dsDNA-MC）用于输入受限的非线性预测控制器设计,提出了基于dsDNA-MC优化的非线性系统预测控制算法。在对单输入单输出非线性系统预测控制分析的基础上,将非线性系统预测控制问题归结为具有输入约束的非线性系统优化问题,并采用dsDNA-MC算法来求解这一问题。仿真结果表明该算法可行、有效。     

粒子群算法量子模型

针对目前粒子群优化(PSO)算法理论基础薄弱,算法本质的分析还未形成体系的问题,从微观的角度出发,以量子力学为基础,提出并建立了粒子群优化算法的量子模型.模型采用无限深方势阱为分析背景,将算法的搜索过程解释为量子状态的转换,并通过模型解释算法执行过程中的内部机制,最后通过实验证明了所提出PSO算法寻优的量子本质.     

基于Lotka-Volterra 模型的双群协同竞争粒子群优化算法

针对粒子群优化算法易出现早熟收敛的问题,提出了基于Lotka-Volterra模型的双群协同竞争粒子群优化算法(LVPSO).LVPSO算法借鉴种群生态学中著名的Lotka-Volterra双群协同竞争模型,讨论了两种种群协同竞争方案,通过群内和群间竞争增加粒子的多样性,提高了种群摆脱局部极值的能力.对5个典型基准测试函数进行优化实验表明,LVPSO在收敛速度和优化精度方面均有良好的表现.     

一种求解多峰函数优化问题的量子行为粒子群算法

介绍了一种利用量子行为粒子群算法（QPSO）求解多峰函数优化问题的方法。为此,在QPSO中引进一种物种形成策略,该方法根据群体微粒的相似度并行地分成子群体。每个子群体是围绕一个群体种子而建立的。对每个子群体通过QPSO算法进行最优搜索,从而保证每个峰值都有同等机会被找到,因此该方法具有良好的局部寻优特性。将基于物种形成的QPSO算法与粒子群算法(PSO)对多峰优化问题的结果进行比较。对几个重要的测试函数进行仿真实验结果证明,基于物种形成的QPSO算法可以尽可能多地找到峰值点,峰值收敛性能优于PSO。     

基于改进PSO算法的过热汽温神经网络预测控制

将改进粒子群优化算法(MPSO)融合到神经网络预测控制中,提出了基于MPSO-RBF混合优化策略的模型预测器,以及基于MPSO算法的非线性优化控制器.针对过热汽温的控制,构造了基十神经网络预测控制的串级控制系统,并就该系统在实现时所涉及到的预测模型、滚动优化算法、反馈校正、仿真参数设置问题等进行了分析,给出了MPSO算法的粒子编码、操作设计和混合优化算法步骤.对某超临界600 MW直流锅炉高温过热器的过热汽温控制,进行了仿真试验,结果表明该方法具有良好的性能指标和应用前景.     

群活性与粒子群优化的稳定性分析

在探讨粒子轨迹的随机过程的基础上,用根轨迹特征值的谱半径来描述粒子群优化的PSO动态系统的稳定性区域;提出并结合实例用群活性刻画了PSO稳定区域中不同参数区间上群行为的动态特征,利用不动点技术通过数值实验描绘出PSO群活性谱及性能图,解释了先前一些文献上提出的典型参数集之所以能够取得满意性能的理由,利用PSO稳定三角中线提出保证PSO收敛性能的参数设置指导策略.     

改进粒子群优化的关联交叉口协调控制

提出了基于改进粒子群优化的关联交叉口协调控制方法。建立了关于排队长度的交通流模型和协调控制目标函数,利用改进粒子群算法对各交叉口绿信比和考虑双向绿波的相位差进行求解,实现了关联交叉口的最优控制。以实际采集的几个关联交叉口的交通数据仿真表明,相比单向绿波控制和感应控制,所提方法可有效减少延误和平均排队长度。     

Position control of electro-hydraulic actuator system using fuzzy logic controller optimized by particle swarm optimization

The position control system of an electro-hydraulic actuator system (EHAS) is investigated in this paper. The EHAS is developed by taking into consideration the nonlinearities of the system: the friction and the internal leakage. A variable load that simulates a realistic load in robotic excavator is taken as the trajectory reference. A method of control strategy that is implemented by employing a fuzzy logic controller (FLC) whose parameters are optimized using particle swarm optimization (PSO) is proposed. The scaling factors of the fuzzy inference system are tuned to obtain the optimal values which yield the best system performance. The simulation results show that the FLC is able to track the trajectory reference accurately for a range of values of orifice opening. Beyond that range, the orifice opening may introduce chattering, which the FLC alone is not sufficient to overcome. The PSO optimized FLC can reduce the chattering significantly. This result justifies the implementation of the proposed method in position control of EHAS.     

A robust evolutionary feedforward active noise control system using Wilcoxon norm and particle swarm optimization algorithm

The conventional filtered-x least mean square (FxLMS) algorithm commonly employed for active noise control (ANC) is sensitive to disturbances acquired by the error microphone and yields poor performance in such scenario. To circumvent this problem, in this paper, a Wilcoxon FxLMS (WFxLMS) algorithm is proposed and used in the design of an efficient ANC which is robust to outliers in the secondary path and immune to burst noise acquired by the error microphone. It is demonstrated through simulation study that under such situation the proposed algorithm outperforms the traditional FxLMS algorithm. A particle swarm optimization (PSO) algorithm based robust ANC system, which does not require the modeling of the secondary path is also derived in the paper. Improved performance of the robust evolutionary ANC system over L2 norm based evolutionary ANC system is also shown.     

Nonlinear model predictive control using trust-region derivative-free optimization

Gradient-based optimization may not be suited if the objective and constraint functions in a nonlinear model predictive control (NMPC) optimization problem are not differentiable. Some well-known derivative-free optimization (DFO)-algorithms are investigated, and a novel warm-start modification to the Wedge DFO-algorithm is proposed. Together with a gradient-based SQP-algorithm these are applied to the NMPC problem and compared in a single-shooting NMPC formulation to a subsea oil–gas separation process. The findings are that DFO is significantly more robust against the numerical issues, compared to a gradient-based SQP tested. Moreover, the warm-start modification reduces the computational complexity.     

Nonlinear model predictive control using deterministic global optimization

This paper presents a Nonlinear Model Predictive Control (NMPC) algorithm utilizing a deterministic global optimization method. Utilizing local techniques on nonlinear nonconvex problems leaves one susceptible to suboptimal solutions at each iteration. In complex problems, local solver reliability is difficult to predict and dependent upon the choice of initial guess. This paper demonstrates the application of a deterministic global solution technique to an example NMPC problem. A terminal state constraint is used in the example case study. In some cases the local solution method becomes infeasible, while the global solution correctly finds the feasible global solution. Increased computational burden is the most significant limitation for global optimization based online control techniques. This paper provides methods for improving the global optimization rates of convergence. This paper also shows that globally optimal NMPC methods can provide benefits over local techniques and can successfully be used for online control.     

Self-adapting control parameters in particle swarm optimization

Particle Swarm Optimization (PSO) is a powerful nature-inspired metaheuristic optimization method. Compared to other methods, PSO can determine the optimal solution in fewer evaluations and generally performs more efficiently and effectively. However, researches show that the PSO method suffers from premature convergence and a dependence on the initial control settings. Due to these shortcomings, the application of PSO may lead to failure in obtaining the global optimal solution. In this work, modifications were performed on the original PSO algorithm to adapt the control parameters to the circumstances of the particles at a specific moment. The proposed method is known as the Unique Adaptive Particle Swarm Optimization (UAPSO). In the developed approach, constraints were handled by forcing the particles to learn from their feasible solutions only. Therefore, the constraint handling technique worked in accord with the adapting scheme to ensure that the particles were adapting to the environment by directing itself to the feasible regions. The performance of UAPSO was verified by a comparative study involving eight benchmark constrained optimization problems and a real-world design problem. The numerical results showed the superiority of UAPSO compared to the selected state-of-the-art metaheuristic methods and PSO variants, its ability in avoiding premature convergence and its consistency and efficiency.