矩阵方程X+A~*X~(-n)A=P的Hermite正定解及其扰动分析

考虑非线性矩阵方程X A~*X~(-n)A=P,其中A是m阶非奇异复矩阵,P是m阶Hermite正定矩阵.本文利用不动点理论讨论了该方程Hermite正定解的存在性及包含区间,给出了极大解的性质及求极大,极小解的迭代算法.研究了极大解的扰动问题,利用微分等方法获得了两个新的一阶扰动界,并给出数值例子对所得结果进行了比较说明.     

矩阵的扰动与广义逆

本文采用文[1]中的术语与记号,个别不同之处则另作说明.文[2]研究了将一奇异方阵扰动到一非奇异方阵的方法,并且给出了用此扰动方法计算方阵的 Moore-Penrose 逆的若干结果.本文研究了将一奇异方阵扰动到非奇异方阵的一般理论,刻划了为得到方阵的某种广义逆而所需的扰动的特征性质;并将这些理论与性质搬到一般的长方矩阵上去,给出了将任意矩阵扰动到非奇异阵的某些一般性结果;最后给出了我们结果的几个应用.     

低秩稀疏矩阵优化问题的模型与算法

低秩稀疏矩阵优化问题是一类带有组合性质的非凸非光滑优化问题.由于零模与秩函数的重要性和特殊性,这类NP-难矩阵优化问题的模型与算法研究在过去十几年里取得了长足发展。本文从稀疏矩阵优化问题、低秩矩阵优化问题、低秩加稀疏矩阵优化问题、以及低秩张量优化问题四个方面来综述其研究现状;其中,对稀疏矩阵优化问题,主要以稀疏逆协方差矩阵估计和列稀疏矩阵优化问题为典例进行概述,而对低秩矩阵优化问题,主要从凸松弛和因子分解法两个角度来概述秩约束优化和秩(正则)极小化问题的模型与算法研究。最后,总结了低秩稀疏矩阵优化研究中的一些关键与挑战问题,并提出了一些可以探讨的问题。     

基于FHM模型的非线性网络化控制系统H_∞优化控制

研究了一类具有随机网络诱导时延,数据包丢失和扰动的网络化控制系统H_∞优化控制问题.针对一类非线性被控对象,基于状态反馈控制器,建立了网络化模糊双曲正切混杂系统模型.根据Lyqpunov稳定性理论,给出了系统可实现γ-次优H_∞镇定的充分条件和控制器优化参数的获取方法.仿真算例利用神经网络算法和线性矩阵不等式工具箱,完成了模型参数的辨识与最优扰动衰减度的求取,证明了分析方法和结果的有效性.     

Hermite矩阵与可对称化矩阵特征值之间的扰动上界

正1引言矩阵特征值的扰动问题,就是研究矩阵元素的改变对矩阵特征值的影响.设矩阵A,B为n阶复矩阵,矩阵B为矩阵A经过扰动之后的矩阵,且λ(A)={λ_i},λ(B))={μ_i},研究矩阵特征值的扰动就是研究λ(A)与λ(B)之间的差距,一般用2范数和Frobenius范数来描述它们之间的差距.矩阵特征值问题是由于处理数据时存在误差而引起的,使得到的特征值往往是经过     

低秩矩阵优化若干新进展

低秩矩阵优化是一类含有秩极小或秩约束的矩阵优化问题,在统计与机器学习、信号与图像处理、通信与量子计算、系统识别与控制、经济与金融等众多学科领域有着广泛应用,是当前最优化及其相关领域的一个重点研究方向.然而,低秩矩阵优化是一个NP-难的非凸非光滑优化问题,其研究成果并非十分丰富,亟待进一步深入研究.主要从理论和算法两个方面总结和评述若干新结果,同时列出相关的重要文献,奉献给读者.     

非光滑参数规划的微分稳定性

稳定性理论是数学规划的重要理论问题之一.主要研究约束集合、扰动函数(最优值函数)、最优解集合与参数扰动之间的关系.稳定性问题的研究也有助于探讨算法收敛性和稳定性.稳定性问题的研究始于70年代.Rockafellar 等人,首先研究了凸规划的稳定性.近些年来才开始研究一般非凸规划的稳定性.Gauvin 等人对非凸规划研究了扰动函数的稳定性与微分稳定性问题,讨论了在一些特殊形式参数扰动的情况下,     

二次规划逆问题的牛顿方法

针对二次规划逆问题,将其表达为带有互补约束的锥约束优化问题.借助于对偶理论,将问题转化为变量更少的线性互补约束非光滑优化问题.通过扰动的方法求解转化后的问题并证明了收敛性.采用非精确牛顿法求解扰动问题,给出了算法的全局收敛性与局部二阶收敛速度.最后通过数值实验验证了该算法的可行性.     

矩阵Frobenius范数不等式

1 引言与引理 矩阵范数与矩阵奇异值问题是数值代数的重要课题,并在矩阵扰动分析,数值计算等分支中起着重要作用.国内外学者对此已作了大量研究.     

带形上近临界随机游动的常返暂留性

本文研究带形上的近临界随机游动,借助游动常返暂留性判别准则的显式表达,通过带扰动的线性差分系统的解的渐近性理论,以及矩阵的范数性质,在扰动矩阵不同的阶的条件下,给出了游动常返暂留性的判别.     

Spectral operators of matrices

Mathematical Programming - The class of matrix optimization problems (MOPs) has been recognized in recent years to be a powerful tool to model many important applications involving structured low...     

Transfer weight functions for injecting problem information in the multi-objective CMA-ES

The covariance matrix adaptation evolution strategy (CMA-ES) is one of the state-of-the-art evolutionary algorithms for optimization problems with continuous representation. It has been extensively applied to single-objective optimization problems, and different variants of CMA-ES have also been proposed for multi-objective optimization problems (MOPs). When applied to MOPs, the traditional steps of CMA-ES have to be modified to accommodate for multiple objectives. This fact is particularly evident when the number of objectives is higher than 3 and, with a high probability, all the solutions produced become non-dominated. An open question is to what extent information about the objective values of the non-dominated solutions can be injected in the CMA-ES model for a more effective search. In this paper, we investigate this general question using several metrics that describe the quality of the solutions already evaluated, different transfer weight functions, and a set of difficult benchmark instances including many-objective problems. We introduce a number of new strategies that modify how the probabilistic model is learned in CMA-ES. By conducting an exhaustive empirical analysis on two difficult benchmarks of many-objective functions we show that the proposed strategies to infuse information about the quality indicators into the learned models can achieve consistent improvements in the quality of the Pareto fronts obtained and enhance the convergence rate of the algorithm. Moreover, we conducted a comparison with a state-of-the-art algorithm from the literature, and achieved competitive results in problems with irregular Pareto fronts.     

A multi-objective memetic algorithm based on decomposition for big optimization problems

When solving multi-objective optimization problems (MOPs) with big data, traditional multi-objective evolutionary algorithms (MOEAs) meet challenges because they demand high computational costs that cannot satisfy the demands of online data processing involving optimization. The gradient heuristic optimization methods show great potential in solving large scale numerical optimization problems with acceptable computational costs. However, some intrinsic limitations make them unsuitable for searching for the Pareto fronts. It is believed that the combination of these two types of methods can deal with big MOPs with less computational cost. The main contribution of this paper is that a multi-objective memetic algorithm based on decomposition for big optimization problems (MOMA/D-BigOpt) is proposed and a gradient-based local search operator is embedded in MOMA/D-BigOpt. In the experiments, MOMA/D-BigOpt is tested on the multi-objective big optimization problems with thousands of variables. We also combine the local search operator with other widely used MOEAs to verify its effectiveness. The experimental results show that the proposed algorithm outperforms MOEAs without the gradient heuristic local search operator.     

The reduction of computation times of upper and lower tolerances for selected combinatorial optimization problems

Due to the growing interest in approximation for multiobjective optimization problems (MOPs), a theoretical framework for defining and classifying sets representing or approximating solution sets for MOPs is developed. The concept of tolerance function is proposed as a tool for modeling representation quality. This notion leads to the extension of the traditional dominance relation to \(t\hbox {-}\)dominance. Two types of sets representing the solution sets are defined: covers and approximations. Their properties are examined in a broader context of multiple solution sets, multiple cones, and multiple quality measures. Applications to complex MOPs are included.     

A fast Pareto genetic algorithm approach for solving expensive multiobjective optimization problems

We present a new multiobjective evolutionary algorithm (MOEA), called fast Pareto genetic algorithm (FastPGA), for the simultaneous optimization of multiple objectives where each solution evaluation is computationally- and/or financially-expensive. This is often the case when there are time or resource constraints involved in finding a solution. FastPGA utilizes a new ranking strategy that utilizes more information about Pareto dominance among solutions and niching relations. New genetic operators are employed to enhance the proposed algorithm’s performance in terms of convergence behavior and computational effort as rapid convergence is of utmost concern and highly desired when solving expensive multiobjective optimization problems (MOPs). Computational results for a number of test problems indicate that FastPGA is a promising approach. FastPGA yields similar performance to that of the improved nondominated sorting genetic algorithm (NSGA-II), a widely-accepted benchmark in the MOEA research community. However, FastPGA outperforms NSGA-II when only a small number of solution evaluations are permitted, as would be the case when solving expensive MOPs.     

An efficient Differential Evolution based algorithm for solving multi-objective optimization problems

In the present study, a modified variant of Differential Evolution (DE) algorithm for solving multi-objective optimization problems is presented. The proposed algorithm, named Multi-Objective Differential Evolution Algorithm (MODEA) utilizes the advantages of Opposition-Based Learning for generating an initial population of potential candidates and the concept of random localization in mutation step. Finally, it introduces a new selection mechanism for generating a well distributed Pareto optimal front. The performance of proposed algorithm is investigated on a set of nine bi-objective and five tri-objective benchmark test functions and the results are compared with some recently modified versions of DE for MOPs and some other Multi Objective Evolutionary Algorithms (MOEAs). The empirical analysis of the numerical results shows the efficiency of the proposed algorithm.     

A new orthogonal evolutionary algorithm based on decomposition for multi-objective optimization

The diversity of solutions is very important for multi-objective evolutionary algorithms to deal with multi-objective optimization problems (MOPs). In order to achieve the goal, a new orthogonal evolutionary algorithm based on objective space decomposition (OEA/D) is proposed in this paper. To be specific, the objective space of an MOP is firstly decomposed into a set of sub-regions via a set of direction vectors, and OEA/D maintains the diversity of solutions by making each sub-region have a solution to the maximum extent. Also, the quantization orthogonal crossover (QOX) is used to enhance the search ability of OEA/D. Experimental studies have been conducted to compare this proposed algorithm with classic MOEA/D, NSGAII, NICA and D2MOPSO. Simulation results on six multi-objective benchmark functions show that the proposed algorithm is able to obtain better diversity and more evenly distributed Pareto fronts than other four algorithms.     

Radial Solutions and Orthogonal Trajectories in Multiobjective Global Optimization

This paper presents a new, ray-oriented method for the global solution of nonscalarized vector optimization problems and a framework for the application of the Karush–Kuhn–Tucker theorem to such problems. Properties of nonlinear multiobjective problems implied by the Karush–Kuhn–Tucker necessary conditions are investigated. The regular case specific to nonscalarized MOPs is singled out when a nonlinear MOP with nonlinearities only in the constraints reduces to a nondegenerate linear system. It is shown that the trajectories of the Lagrange multipliers corresponding to the components of the vector cost function are orthogonal to the corresponding trajectories of the vector deviations in the balance space (to the balance set for Pareto solutions). Illustrative examples are presented.     

A multiobjective immune algorithm based on a multiple-affinity model

This paper presents a new multiobjective immune algorithm based on a multiple-affinity model inspired by immune system (MAM-MOIA). The multiple-affinity model builds the relationship model among main entities and concepts in multiobjective problems (MOPs) and multiobjective evolutionary algorithms (MOEAs), including feasible solution, variable space, objective space, Pareto-optimal set, ranking and crowding distance. In the model, immune operators including clonal proliferation, hypermutation and immune suppression are designed to proliferate superior antibodies and suppress the inferiors. MAM-MOIA is compared with NSGA-II, SPEA2 and NNIA in solving the ZDT and DTLZ standard test problems. The experimental study based on three performance metrics including coverage of two sets, convergence and spacing proves that MAM-MOIA is effective for solving MOPs.     

Evaluating selection methods on hyper-heuristic multi-objective particle swarm optimization

Multi-objective particle swarm optimization (MOPSO) is a promising meta-heuristic to solve multi-objective problems (MOPs). Previous works have shown that selecting a proper combination of leader and archiving methods, which is a challenging task, improves the search ability of the algorithm. A previous study has employed a simple hyper-heuristic to select these components, obtaining good results. In this research, an analysis is made to verify if using more advanced heuristic selection methods improves the search ability of the algorithm. Empirical studies are conducted to investigate this hypothesis. In these studies, first, four heuristic selection methods are compared: a choice function, a multi-armed bandit, a random one, and the previously proposed roulette wheel. A second study is made to identify if it is best to adapt only the leader method, the archiving method, or both simultaneously. Moreover, the influence of the interval used to replace the low-level heuristic is analyzed. At last, a final study compares the best variant to a hyper-heuristic framework that combines a Multi-Armed Bandit algorithm into the multi-objective optimization based on decomposition with dynamical resource allocation (MOEA/D-DRA) and a state-of-the-art MOPSO. Our results indicate that the resulting algorithm outperforms the hyper-heuristic framework in most of the problems investigated. Moreover, it achieves competitive results compared to a state-of-the-art MOPSO.