GSVM优化问题的一种新的光滑函数法

提出求解广义支撑向量机(GSVM)优化问题的一种新的光滑函数法,克服了已有算法收敛速度慢且计算结构复杂的缺陷。首先利用最优化理论的KKT互补条件,将GSVM转化为无约束优化问题,然后给出了基于Newton型迭代的光滑函数的迭代方法。给出了这种光滑函数的有关性质、迭代算法的迭代格式及其收敛性。通过理论分析及数值实验证明了该算法对初始点不敏感,且收敛速度快、数值稳定。从而验证了算法的可行性和有效性。     

Continuity of solution mappings for parametric generalized set-valued weak vector equilibrium problems

This paper obtains some stability results for parametric generalized set-valued weak vector equilibrium problem. Under new assumptions, which do not contain any information about solution mappings, the authors establish the continuity of the solution mapping to a parametric generalized set-valued weak vector equilibrium problem without monotonicity. These results extend and improve some results in the literature. Some examples are given to illustrate the results.     

Solving large-scale multiclass learning problems via an efficient support vector classifier

1. INTRODUCTION In recent years, a new type of classifier, support vector machines[1~2], is receiving adoption increasingly as a state-of-the-art tool to solve knowledge discovery pro- blems. SVMs are based on the statistical learning the- ory of Vapnik[1] and quadratic programming optimiz- ation. Support vector machines (SVMs) are initially designed for binary classification problem. How to effectively extend them for multiclass classification is still an ongoing research topic. Curr…     

Data-driven evolutionary sampling optimization for expensive problems

Surrogate models have shown to be effective in as-sisting evolutionary algorithms (EAs) for solving computationally expensive complex optimization problems. How...     

ON VECTOR NETWORK EQUILIBRIUM PROBLEMS

1 Introduction The earliest network equilibrium model was proposed by Wardrop (1952) for a transportation network Since then, many other equilibrium models have also been proposed in the economics literature (see Nagurney 1993) Until only recently, all these equilibrium models are based on single cost or utility function Recently, equilibrium models based on multicriteria consideration or vector-valued cost functions have been proposed In Chen and Yen (1993), a multicriteria traffic equilibr…     

无圈超图对策上的有效平均树解

在超图对策中,平均树解(简称AT值)是一类重要的分支有效解.本文在大联盟能够形成的假设下,基于单而芳等(2017)和Béal等(2018)在无圈图对策上提出的有效平均树解,进一步提出了无圈超图对策上基于平均树解的一类有效解.该有效解先给每个参与者进行AT值分配,再将大联盟收益超出分支收益之和的剩余部分平均分配给每个参与者.首先,证明了该有效解可以由有效性、分支公平性和分支内盈余等分性这三个公理所唯一刻画.其次,通过算例分析了该有效解.通过与其他值比较发现,对处于关键位置的参与者,在该解下会获得更多的收益.     

求解大规模优化问题的改进狼群算法

针对狼群算法(wolf pack algorithm,WPA)求解大规模问题时存在解精度低、收敛速度慢和易陷入局部最优的缺点,提出了一种改进的狼群算法(improved wolf pack algorithm,IWPA).采用深度神经网络进行初始化狼群个体提高种群多样性;借助遗传算法挑选首狼提高算法寻优能力;设计距离优化因子以协同狼群算法个体的探索和开发能力;构建尺度系数改进围捕行为避免算法陷入局部最优,减少运行时间.选取18个大规模(100维,200维,500维和1000维)标准测试函数进行性能对比,结果表明,IWPA算法在求解精度和收敛速度上优于其他对比算法.     

Henig globally efficiency for set-valued optimization and vector variational inequality

This paper deals with Henig globally efficiency in vector optimization involving generalized cone-preinvex set-valued mapping. Some properties of generalized cone-preinvex set-valued map are derived. It also disclose the closed relationships between Henig globally efficiency of generalized conepreinvex set-valued optimization problem and Henig globally efficiency of a kind of vector variational inequality.     

Novel electromagnetism-like mechanism method for multiobjective optimization problems

As a new-style stochastic algorithm, the electromagnetism-like mechanism(EM) method gains more and more attention from many researchers in recent years. A novel model based on EM(NMEM) for multiobjective optimization problems is proposed, which regards the charge of all particles as the constraints in the current population and the measure of the uniformity of non-dominated solutions as the objective function. The charge of the particle is evaluated based on the dominated concept, and its magnitude determines the direction of a force between two particles. Numerical studies are carried out on six complex test functions and the experimental results demonstrate that the proposed NMEM algorithm is a very robust method for solving the multiobjective optimization problems.     

A general iterative method of fixed points for equilibrium problems and optimization problems

The purpose of this paper is to present a general iterative scheme as below:

矢量波动方程的直接差分解法

对电磁场波动方程进行了分析,研究了一种基于电场波动方程的时域数值方法———矢量波动方程的直接差分解法,给出了电场和磁场的计算公式。采用线性插值方法获得了导体表面法向电场的计算公式,给出了算法的稳定性条件。与时域有限差分法相比,由于该方法中电场分量的迭代更新无需磁场参与,从而使算法的复杂性大大降低,是一种具有良好应用前景的电磁场时域数值分析方法,最后通过一个数值计算实例证明了该算法的有效性。     

求解约束优化问题的改进粒子群算法

针对高维复杂约束优化问题,提出了一种基于平滑技术和一维搜索的粒子群算法(NPSO)。该算法使粒子的飞行无记忆性,结合平滑函数和一维搜索重新生成停止进化粒子的位置,增强了在最优点附近的局部搜索能力;定义了不可行度阈值,利用此定义给出了新的粒子比较准则,该准则可以保留一部分性能较优的不可行解微粒,使微粒能快速的找到位于约束边界或附近的最优解;最后,为了扩大粒子的搜索范围,引进柯西变异算子。仿真结果表明,对于复杂约束优化问题,算法寻优性能优良,特别是对于超高维约束优化问题,该算法获得了更高精度的解。     

一种求解多峰优化问题的改进Species粒子群算法

为了改善粒子群算法求解多峰优化问题的能力,提出了一种新的利用粒子在种群中序号来构造Species的方法,同时还设计了适应性局域搜索算子和触发式重新初始化策略分别用于增强算法的开发能力和探索能力.通过对一组标准多峰测试函数的仿真实验,表明所提出的算法能够有效的求解多峰优化问题.     

Parameter selection of support vector machine for function approximation based on chaos optimization

The support vector machine （SVM） is a novel machine learning method, which has the ability to approximate nonlinear functions with arbitrary accuracy. Setting parameters well is very crucial for SVM learning results and generalization ability, and now there is no systematic, general method for parameter selection. In this article, the SVM parameter selection for function approximation is regarded as a compound optimization problem and a mutative scale chaos optimization algorithm is employed to search for optimal paraxneter values. The chaos optimization algorithm is an effective way for global optimal and the mutative scale chaos algorithm could improve the search efficiency and accuracy. Several simulation examples show the sensitivity of the SVM parameters and demonstrate the superiority of this proposed method for nonlinear function approximation.     

约束多目标优化问题的进化算法及其收敛性

提出了带约束多目标优化问题的一种新解法。首先定义了个体的序值和个体的约束度,利用这两个定义给出了一种新的适应度函数和开关选择算子,从而对种群中的个体进行评估或排序时无需特别关心个体是否可行,避免了罚函数选择参数的困难。用概率论有关理论证明了算法的收敛性。用标准的Benchmark函数进行了仿真实验,仿真结果表明,新算法对约束多目标优化问题的求解是有效的。     

Some inverse optimization problems on network

In this paper, the authors consider some inverse problems on network, such as the inverse transport problems with gains (IGTP) and the inverse linear fractional minimum cost flow problem (IFFP). Firstly, the authors give the mathematics model of (IGTP) and an efficient method of solving it under l ₁ norm; Secondly, taking advantage of the optimality conditions, the authors consider the (IFFP) and give a simple method of solving it. Finally, an numerical example test is also developed.     

Hierarchical control based on Hopfield network for nonseparable opti mization problems

1 .INTRODUCTIONConsider thefollowing complexsystemthat comprisesNinterconnected subsystems , the mathematicalmodel of theith subsystemisyi=fi(ci,ui,zi)gi(ci,ui,yi) ≥0ui= Hiywhereci, ui, yiandziare theithsubsystemcontrolinput ,interactioninput ,output and disturbance vec-tor respectively,andfiis the mathematics model oftheithsubsystem.It is assumedthat external distur-bancezwhich affects the systemis constant in theconsidered ti me interval of control and hence will beleft out of the syst…     

Generalized invexity-type conditions in constrained optimization

This paper defines a new class of generalized type I functions, and obtains Kuhn-Tucker necessary and sufficient conditions and duality results for constrained optimization problems in the presence of the aforesaid weaker assumptions on the objective and constraint functions involved in the problem.     

一类非线性大系统优化问题的逼近算法

针对一类约束函数是变量可分离的非线性大系统优化问题,本文给出一种基于逐次逼近算法的熵函数法,对每一个子问题,该方法可以通过解一个可微的无约束极小值问题,一次性地获得其ε－最优解,避免了序列极小化过程,并且可以通过参数的选承控制解的误差,初步的数值试验表明,对于该类非线性大系统优化问题,本算法有良好的数值表现。