一种新的时间数据序列缓冲技术及其应用研究

以排除扰动因素对系统特征时间序列的冲击,还系统真实面貌,增强系统特征时间序列预测建模精度为目的,根据新信息优先利用原理,在对序列缓冲算子理论和已有序列弱化缓冲算子研究的基础上,构造了一类新的时间序列弱化缓冲算子,并与已有的系列弱化算子进行了建模精度比较。研究结果表明这类新的序列弱化缓冲算子能显著提高小样本时间序列的预测精度。     

基于单调函数的新弱化缓冲算子研究

在灰色系统缓冲算子公理体系下,构造了2类新弱化缓冲算子,并将其与党氏弱化缓冲算子进行比较,论证了党氏弱化缓冲算子为新算子的特例,从而大大地拓广了弱化缓冲算子的应用范围.对序列前一部分增长(衰减)速度过快,而后一部分增长(衰减)速度过慢的冲击扰动系统数据序列,在建模预测过程中常常出现的定量预测结果与定性分析结论不符的问题,提供了多种解决方案,首次将缓冲算子的构造与函数联系起来,从而为缓冲算子的构造开辟了新方向.     

平滑变权缓冲算子构造及其性质

针对变权缓冲算子信息利用不充分以及权重选择问题, 提出一类新的平滑变权缓冲算子. 研究了该缓冲算子的性质, 证明了平滑变权缓冲算子对序列具有弱化作用并能够提升序列光滑性, 得出了平滑变权缓冲算子调节度的递推不等式; 通过多目标优化方法来确定可变权重取值, 构造可变权重的优化目标函数, 并结合遗传算法来确定权重的最优取值. 实例分析表明, 所提出的平滑变权缓冲算子能够有效提高建模精度.

     

一类新的弱化缓冲算子的构造及其应用

利用灰色系统理论,在对缓冲算子和已有弱化缓冲算子研究的基础上,构造一类新的弱化缓冲算子,并研究其一些特性及内在关系,有效解决了冲击扰动数据序列在建模预测过程中经常出现的定量预测结果与定性分析结论不符的问题.实例分析结果表明,这类新的弱化缓冲算子能显著提高数据预测模型的预测精度.     

变权缓冲算子及其作用强度的研究

针对传统缓冲算子不能实现作用强度的微调,从而导致缓冲作用效果过强或过弱的问题,构造了变权弱化缓冲算子和变权强化缓冲算子.研究了缓冲算子调节度与可变权重之间的关系,并用遗传算法探讨该类缓冲算子的优化问题.研究结果表明,可变权重在功能上类似于高阶作用算子,但控制缓冲算子作用强度的灵活性则明显优于高阶缓冲算子.最后以我国能源消费总量的预测问题为例,验证了变权缓冲算子的有效性和优越性.     

关于新的弱化缓冲算子的研究及其应用

在灰色系统理论缓冲算子公理体系下,构造出一类新的弱化缓冲算子.研究其一些特性和内在联系,可有效地解决冲击扰动数据序列在建模预测过程中经常出现的定量预测结果与定性分析结论不符的问题.实例分析结果表明,该类算子是有效而实用的.     

基于不动点的新强化缓冲算子及其性质研究

在灰色系统缓冲算于公理体系下,利用反函数定理构造了两类新强化缓冲算子,并研究其特性及各种强化缓冲算子之间的内在关系,有效地解决了冲击扰动系统数据序列建模预测过程中常常出现的定量预测结果与定性分析结论不符的问题.将缓冲算子的构造与函数联系起来,从而推广了缓冲算子的类型.     

基于一类新的强化缓冲算子的GM(1,1)预测精度研究

根据灰色系统理论中新息优先利用原理,在对缓冲算子和已有强化缓冲算子研究的基础上,构造了一类新的强化缓冲算子,有效解决了冲击扰动数据序列在建模预测过程中常常出现的定量预测结果与定性分析结论不符的问题.实例分析结果表明,这类新的强化缓冲算子能显著提高GM(1,1)模型的预测精度.     

基于单调函数的新弱化缓冲算子研究

在灰色系统缓冲算子公理体系下,构造了2类新弱化缓冲算子,并将其与党氏弱化缓冲算子进行比较,论证了党氏弱化缓冲算子为新算子的特例,从而大大地拓广了弱化缓冲算子的应用范围.对序列前一部分增长(衰减)速度过快,而后一部分增长(衰减)速度过慢的冲击扰动系统数据序列,在建模预测过程中常常出现的定量预测结果与定性分析结论不符的问题,提供了多种解决方案,首次将缓冲算子的构造与函数联系起来,从而为缓冲算子的构造开辟了新方向.

     

基于不动点的强化缓冲算子序列及其应用

在灰色系统理论缓冲算子公理体系下,构造了4类强化缓冲算子序列,整合了已有的实用强化缓冲算子,并对它们的缓冲作用进行比较.实例验证了这类算子序列的有效性和实用性,解决了冲击扰动系统行为数据在建模预测过程中出现的定性分析与定量预测不符的问题.     

几类强弱缓冲算子的构造方法及其内在联系

基于最基本的缓冲算子x(k)d=∑ni=kωix(i)/∑ni=kωi变形,获得了几类范围大大拓宽的实用缓冲算子的构造方法,并揭示了这几类缓冲算子的内在联系.同时揭示了强化缓冲与弱化缓冲的对应关系,即指数γ由正到0,由0到负的变化,导致缓冲算子由弱(强)化到不变化,由不变化到强(弱)化的过程,揭示了由量变到质变的过程.研究表明,现有的大量缓冲算子都是所构造的缓冲算子的特例,此外还包括许多与常见文献中类型迥然不同的新算子.     

Using the Fractional Order Method to Generalize Strengthening Buffer Operator and Weakening Buffer Operator

To reveal the relationship between a weakening buffer operator and strengthening buffer operator, the traditional integer order buffer operator is extended to one that is fractional order. Fractional order buffer operator not only can generalize the weakening buffer operator and the strengthening buffer operator, but also results in small adjustments of the buffer effect. The effectiveness of the grey model (GM(1, 1)) with the fractional order buffer operator is validated by six cases.      

求解加权圆集布局问题的快速启发式搜索算法

加权圆集布局问题是基于性能驱动的一类布局问题,由于其NP-hard属性,难以在多项式时间内求解,提出一种快速启发式搜索算法。权矩阵的行向量1范数作为首次赌轮选择圆的启发信息,依次以权矩阵的当前行（其行号等于当前选择圆的序号）元素作为下次赌轮选择的启发信息,利用图形学理论给出低计算复杂度的定位规则,进而基于该定序定位规则提出一种启发式搜索算法,以求得该问题的最优解。数值实验表明,该算法的性能优于已有算法。     

带扰动和变异因子的改进禁忌搜索算法求解贴片机贴装过程优化

以多贴装头拱架式贴片机为研究对象,利用带扰动和变异因子的改进禁忌搜索算法,实现贴片机贴装过程优化.算法在传统禁忌搜索算法的基础上,利用以长期记忆为基础的多元化扰动策略和块变异算子来扩大贴片机贴装顺序优化搜索空间,并结合局部下降搜索策略优化喂料器分配,最终实现贴片机贴装整体优化.仿真实验表明,改进算法能快速有效地获得较好的贴片机贴装优化解.与其他文献中的算法相比,所提出的算法在求解质量和求解速度上有较大的优越性.     

Optimal Design of a 6-DOF Parallel Manipulator Using Particle Swarm Optimization

Performance indices of parallel manipulators (PMs) vary widely with the variation of geometric properties. Improvement of one parameter often leads to worsen the other parameters. Therefore, getting into an optimum design for the PMs has been subject of much recent research. In this paper, we optimize three performance parameters of a PM simultaneously including workspace, condition number, and stiffness. In addition, a new performance index is introduced for stiffness evaluation of the PMs. The index is invariant under similarities. Because of complexity of cost function and number of variables, choosing an optimization method that can converge to the optimum point is very important. We select particle swarm optimization (PSO) method and show that this algorithm is perfect for performance optimization of PMs. Furthermore, we propose a new subroutine added to PSO algorithm to improve its convergence.     

Two new meta-heuristics for a bi-objective supply chain scheduling problem in flow-shop environment

In this study, an integrated multi-objective production-distribution flow-shop scheduling problem will be taken into consideration with respect to two objective functions. The first objective function aims to minimize total weighted tardiness and make-span and the second objective function aims to minimize the summation of total weighted earliness, total weighted number of tardy jobs, inventory costs and total delivery costs. Firstly, a mathematical model is proposed for this problem. After that, two new meta-heuristic algorithms are developed in order to solve the problem. The first algorithm (HCMOPSO), is a multi-objective particle swarm optimization combined with a heuristic mutation operator, Gaussian membership function and a chaotic sequence and the second algorithm (HBNSGA-II), is a non-dominated sorting genetic algorithm II with a heuristic criterion for generation of initial population and a heuristic crossover operator. The proposed HCMOPSO and HBNSGA-II are tested and compared with a Non-dominated Sorting Genetic Algorithm II (NSGA-II), a Multi-Objective Particle Swarm Optimization (MOPSO) and two state-of-the-art algorithms from recent researches, by means of several comparing criteria. The computational experiments demonstrate the outperformance of the proposed HCMOPSO and HBNSGA-II.     

基于变异的Bayesian优化算法

将变异算子与Bayesian优化算法相结合,提出了一种基于变异的Bayesian优化算法。在算法中设计了一个种群多样性函数,通过此函数引入变异算子,目的是利用变异算子的邻域搜索能力,保持种群多样性,将贝叶斯概率模型提取的全局信息与变异算子的局部信息联系起来,避免陷入局部最优。仿真研究表明基于变异的Bayesian优化算法的寻优能力比Bayesian优化算法更强。     

A consensus model for group decision making under trapezoidal fuzzy numbers environment

Natural linguistic terms can preferably express the opinions of decision makers in complicated decision environment. Group decision making with multiplicative trapezoidal fuzzy preference relations transforming from natural linguistic terms attracts the attention of researchers for its important research significant. The developed approach is based on consensus improving process by using a new similarity measure and a trapezoidal fuzzy power ordered weighted geometric averaging (TFPOWGA) operator. In order to introduce this approach, firstly, trapezoidal fuzzy power geometric averaging operator and TFPOWGA operator are presented to aggregate trapezoidal fuzzy numbers (TFNs). Secondly, a new similarity measure for TFNs is introduced by combining centroids and areas of TFNs. We further propose a new consensus improving algorithm that consists of consensus measure and a dynamic feedback mechanism containing a multi-objective optimization model and some indirect rules. And then a selection stage is described to rank the alternatives. At last, an example is implemented to demonstrate effectiveness of the approach.