一种基于加权有向拓扑的改进粒子群算法

研究粒子群优化算法（PSO）的拓扑结构和信息流动，以提高算法性能是PSO的一个有意义的研究方向。RuiMendes等人提出的全联通型算法（FIPSO），其拓扑结构本质上是加权无向图，两个邻接点之间的相互影响是对等的，与社会人际网络的真实情况不符。提出了一种改进型算法，重新构造了加权函数，体现了粒子之间影响的不平衡性。仿真结果显示：该改进算法对收敛速度和稳定性均有非常好的改善。     

层次环形拓扑结构的动态粒子群算法

粒子群算法（PSO）的拓扑结构决定粒子之间的信息交互方式,是影响算法性能的关键因素。为提高算法性能,提出了一种层次环形拓扑结构的动态粒子群算法（HRPSO）,粒子组成的环被分配在规则树中,算法运行时,环在层次中动态移动。通过6个标准测试函数优化,比较了HRPSO与几种基准算法的性能,实验结果证明HRPSO在精确性和稳定性上具有优势。     

混合拓扑结构的粒子群算法及其在测试数据生成中的应用研究

粒子群算法(PSO)的拓扑结构是影响算法性能的关键因素,为了从根源上避免粒子群算法易陷入局部极值及早熟收敛等问题,提出一种混合拓扑结构的粒子群优化算法(MPSO)并将其应用于软件结构测试数据的自动生成中。通过不同邻域拓扑结构对算法性能影响的分析,采用一种全局寻优和局部寻优相结合的混合粒子群优化算法。通过观察粒子群的多样性反馈信息,对每一代种群粒子以进化时选择全局拓扑结构模型(GPSO)或局部拓扑结构模型(LPSO)的方法进行。实验结果表明,MPSO使得种群的多样性得到保证,避免了粒子群陷入局部极值,提高了算法的收敛速度。     

基于拓扑结构与粒子变异改进的粒子群优化算法

为使粒子群优化算法(PSO)优化过程的多样性与收敛性得到合理解决,以提高算法优化性能,基于种群拓扑结构与粒子变异提出两种粒子群改进算法RSMPSO和RVMPSO.改进算法将具有信息定向流动的闭环拓扑结构与星型拓扑结构或四边形拓扑结构相结合,促使粒子在前期寻优过程中具有较高的多样性,保证搜索的广度,而在后期满足粒子群的整体收敛性,保证寻优的精度.同时,将布谷鸟搜索算法(CS)中的偏好随机游走变异策略引入改进算法中,增强粒子跳出局部最优的能力.对标准测试函数的仿真实验表明,所改进的PSO算法与其他6个对比算法相比不仅操作简单,优化精度高,而且在算法收敛性及稳健性方面都有着更出色的表现.     

电力系统无功优化中PSO算法的改进研究

针对电力系统无功优化中的PSO算法的特点,采用的信息拓扑结构为环形结构,对PSO算法中的变异算子进行研究.针对环形拓扑结构的PSO算法,其后期收敛精度差是一个常见问题,提出了一种称之为"球面变异"的变异算子,充分利用粒子群迭代后期种群的信息,对变异的方向与速度进行引导,进而建立了变异算子与当代种群适应度之间的关系,明显地提高了算法收敛速度与精度.最后,对陷入局部收敛等问题进行相应的改良,诸如无法达到最优解等问题.使用IEEE14节点系统作为算例进行测试,结果达到优良.     

一种改进的自适应邻域粒子群优化算法

在对粒子群优化(PSO)算法进行深入分析的基础上,建立了自适应邻域更新机制,再对惯性权重更新机制进行自适应化,分别从拓扑邻域结构和惯性权重两个角度对局部版PSO算法进行了改进,提出了一种实用、高效的自适应邻域粒子群优化算法,经7个标准测试函数验证,该算法具有较高效率和精度。     

动态邻域混合粒子群优化算法

粒子群优化(PSO)算法对于多峰搜索问题一直存在早熟收敛问题。为在增强PSO算法全局搜索能力的同时提高收敛速度,提出一种动态邻域混合粒子群优化算法DNH_PSO,采用PSO局部模型,将随机拓扑和冯诺依曼拓扑相结合形成动态邻域,提高算法的全局搜索能力,为增强算法的局部搜索能力并加快收敛速度,使用粒子邻域全面学习策略,将拟牛顿法引入算法中。与其他PSO实验对比分析表明,该算法对于多峰搜索问题具有较好的全局收敛性。     

一种基于有向动态网络拓扑的粒子群优化算法

该文提出了一种改进的PSO算法PSO-DSF。引进有向类无标度网作为粒子群寻优的拓扑结构,提出作为粒子邻域拓扑的有向网络动态变化机制,使有向网络在出度服从幂律分布的条件下动态变化,从而提高算法的多样性,避免过早陷入局部最优的情况。通过函数测试,证实了该改进方案的有效性。     

一种具有初始化功能的自适应惯性权重粒子群算法

提出了一种改进的具有初始化功能的自适应惯性权重粒子群优化(PSO)算法.该算法首先引入自适应惯性权重策略均衡全局和局部搜索能力,并针对运行过程中出现停滞现象的粒子群,围绕其加权重心位置重新初始化,引导粒子突破了局部极值的限制,提高了算法的收敛速度.最后,将此算法、PSO算法及惯性权重线性递减的PSO(LDW-PSO)算法进行了比较.实验结果表明,该算法不仅有效地增强了粒子突破局部极值的能力,而且算法的收敛速度和稳定性也有了一定的提高.     

基于动态邻居拓扑结构的PSO算法

粒子群优化(PSO)算法在求解复杂的多峰问题时极易陷入局部最优解,通过分析种群多样性与局部最优解间的关系,提出一种基于动态邻居拓扑结构的粒子群算法。该算法在运行过程中,每间隔若干代,根据粒子间的距离更新每个粒子的邻居,该策略增加种群的多样性,进而提升粒子跳出局部最优解的能力。实验结果表明,该算法比其他PSO算法具有更好的性能。     

Task Scheduling in Distributed Systems Using Heap Intelligent Discrete Particle Swarm Optimization

The optimal mapping of tasks to the processors is one of the challenging issues in heterogeneous computing systems. This article presents a task scheduling problem in distributed systems using discrete particle swarm optimization (DPSO) algorithm with various neighborhood topologies. The DPSO is a recent metaheuristic population‐based algorithm. In DPSO, the set of particles in a swarm flies through the N‐dimensional search space by learning from both the personal best position and a neighborhood best position. Each particle inside the swarm belongs to a specific topology for communicating with neighboring particles in the swarm. The neighborhood topology affects the performance of DPSO significantly, because it determines the rate at which information transmits through the swarm. The proposed DPSO algorithm works on dynamic topology that is binary heap tree for communication between the particles in the swarm. The performance of the proposed topology is compared with other topologies such as star, ring, fully connected, binary tree, and Von Neumann. The three well‐known performance measures such as Makespan, mean flow time, and reliability cost are used for the comparison of the proposed topology with other neighborhood topologies. Computational simulation results indicate that the performance of DPSO algorithm has shown significant improvement with binary heap tree topology used for communication among the particles in the swarm.     

无线传感器网络拓扑控制算法研究进展

对于无线传感器网络而言,拓扑控制是一个基本问题,对网络性能的影响很大,其目标是用最小的能量维持网络拓扑。文章较为详细地介绍了现有的无线传感器网络拓扑控制算法,并分析了各种控制算法的优缺点,探讨了拓扑控制算法今后的重点研究方向。     

Dynamic small world particle swarm optimizer for function optimization

Performance of particle swarm optimization technique is highly influenced by the population topology. It determines the way in which particles communicate and share information within a swarm. If path length is too small, it implies that a particle communicates with other particles in its close proximity leading to exploitation. On the contrary, if path length is large then the particle interacts with other remote particles leading to exploration. There needs to be a balance between exploration and exploitation and Small world network fits to this need of ours. In this paper, dynamic small world network has been proposed with the objective to have a balanced trade-off between exploration and exploitation. In order to make learning process dynamic linearly decreasing inertia weight has been employed. Experimental study is performed on a set of 23 test functions using different performance evaluation measures. Results obtained are compared with other state of the art techniques demonstrating the effectiveness of the proposed approach.     

A spatiotemporal-distributed deep-learning framework for KPI estimation of chemical processes with cascaded reactors

Online and accurate estimation of key performance indicators (KPI) is the foundation for operational optimization of a chemical process. However, a chemical process usually consists of multiple reactors, and the factors influencing KPI are spatially distributed in the long process flow. In addition, due to the distinct time lags between KPI and each reactor, temporal relationships among KPI and its influence factors are a mixture of short-term and long-term relationships. In this regard, a deep distributed KPI estimator with a self-attention mechanism is proposed in this paper. First, considering the process topology, a cascaded long short-term memory network is developed to simulate the process topology and capture the short-term effects. Then, to extract the long-term dependencies, a de-noise self-attention layer is employed to model interactions of all the influence factors explicitly and dynamically. Lastly, the proposed method is compared with typical state-of-the-art methods using real industrial data. The comparison results illustrate the performance and effectiveness of the proposed KPI estimation method.     

Performance-oriented communication topology design for distributed control of interconnected systems

Communication networks provide a larger flexibility for the control design of interconnected systems by allowing the information exchange between the local controllers of the subsystems which can be used to improve the overall system performance. However, the interconnected systems may become unstable due to permanent communication link failures. This article presents a novel two-layer control architecture that allows to jointly improve the system performance which is the decay rate and guarantee the stability of the interconnected system under permanent communication link failures. As a novelty, the design of communication topology between the local controllers is also taken into account. On the other hand, it is still not well understood how significant the role of each possible communication link is in improving the system performance. Another novelty of this article is thus to propose a method based on eigenvalue sensitivity analysis in order to characterize the influence of each possible communication link in improving the performance of the overall system. In addition, for a special class of systems and physical interconnection topology, explicit solutions on communication topology design are derived for the first time. The solutions provide some insights into how the heterogeneity of the subsystem local dynamics, the strength of interconnection and the size of the network affect the optimal communication topology.     

基于InfiniBand的多链路mesh/torus大规模并行系统互连网络

在大规模并行系统中,系统级互连网络的设计至关重要.InfiniBand作为一种高性能交换式网络被广泛应用于大规模并行处理系统中.mesh/torus拓扑结构相较于目前普遍应用于InfiniBand网络的胖树拓扑结构拥有更好的性能与可扩展性.尽管如此,研究发现,用传统的mesh/torus拓扑结构构建InfiniBand互连网络存在诸多问题.分析了传统网络拓扑结构的缺陷,并提出了一种基于InfiniBand的多链路mesh/torus互连网络.这种改进型的拓扑结构通过充分利用交换机间的多链路可以获得比传统mesh/torus网络更高的带宽.另外,同时给出了与该网络拓扑结构相配套的高效路由算法.最后,通过网络仿真技术对提出的算法进行了评估,实验结果显示提出的路由算法相较于其他路由算法拥有更好的性能与可扩展性.     

Frequency-domain Tuning of Robust Fixed-structure Controllers via Quantum-behaved Particle Swarm Optimizer with Cyclic Neighborhood Topology

This paper presents a constrained particle swarm optimization (PSO) algorithm with a cyclic neighborhood topology inspired by the quantum behavior of particles, and describes its application to the frequency-domain tuning of robust fixed-structure controllers. Two main methodologies for improving the exploration and exploitation performance of the PSO framework are described. First, a PSO scheme with a neighborhood structure based on a cyclic network topology is presented. This scheme enhances the exploration ability of the swarm and effectively reduces the probability of premature convergence to local optima. Second, the above PSO scheme is hybridized using a distributed quantum-principle-based offspring creation mechanism. Such a hybridized PSO framework enables neighboring particles to concentrate the search around the region covered by those particles to refine the candidate solution. A frequency-domain tuning method for fixed-structure controllers is then demonstrated. This method guarantees certain preassigned performance specifications based on the developed PSO technique. A typical numerical example is considered, and the results clearly demonstrate that the proposed PSO scheme provides a novel and powerful impetus with remarkable reliability for robust fixed-structure controller syntheses. Further, an experiment was conducted on a magnetic levitation system to compare the proposed strategy with a well-known frequency-domain tuning method implemented in the MATLAB tool for Structured H_∞ Synthesis. The comparative experimental results validate the effectiveness of the proposed tuning strategy in practical applications.     

基于动态邻域的粒子群算法的研究

研究了邻域拓扑结构对粒子群算法性能的影响。设计了两种动态邻域生成策略,并基于一组具有代表性的测试函数,对两种典型的算法模型——标准的粒子群算法（CPSO）和充分联系的粒子群算法（FIPS）进行实验。实验结果表明,不同的邻域拓扑结构和不同的算法模型都能够影响粒子群算法的性能。     

基于网络邻域拓扑的粒子群优化算法

探讨类无标度网、全局耦合网、环形网、随机网、星形网等邻域拓扑结构对粒子群优化算法寻优效果的影响。理论分析与实验结果显示,以类无标度网作为邻域拓扑结构的粒子群优化算法在误差范围内的寻优效果最好,收敛速度最快,可以较好地避免陷入局部最优,且网络平均度对粒子群优化算法的寻优效果有一定的影响。