一个用次最优化方法解线性约束凸规划的超线性收敛算法

对于带有线性约束的非线性规划的求解问题已有很多算法.其中文献[1,2]将变尺度法分别与既约梯度法、投影梯度法结合,在一定的假设条件下给出了两种超线性收敛的算法;文献[3]处理了退化问题.Zangwill 提出了用求某些流形上的次最优来求解原线性约束凸规划的方法,即将原规划问题的求解问题转化为一系列的求解线性等式约束的子问题,以图最后找到原问题的最优解所在的流形并解之.这种做法使问题变得简单有其实用价值.文献[5]给出了 Zangwill 算法的改进,讨论了退化问题,但[5]总是假定可     

不等式约束二次规划的一新算法

文献[1]提出了一般等式约束非线性规划问题一种求解途径．文献[2]应用这一途径给出了等式约束二次规划问题的一种算法，本文在文献[1]和[2]的基础上对不等式约束二次规划问题提出了一种新算法．     

求解凸二次规划问题的势下降内点算法

1　引　言二次规划问题的求解是数学规划和工业应用等领域的一个重要课题 ,同时也是解一般非线性规划问题的序列二次规划算法的关键 .求解二次规划问题的早期技术是利用线性规划问题的单纯形方法求解二次规划问题的 KKT最优性必要条件[1 ] .这类算法比较直观 ,但在处理不等式约束时 ,松弛变量的引进很容易导致求解过程的明显减慢 .有效集策略是求解二次规划问题的另一类主要技术 .这类方法一般都是稳定的 ,但随着问题中大量不等式约束的出现 ,其收敛速度将越来越低[2 ] .简约空间技术将所求问题的 Hessian阵投影到自由变量所在的子空间中 …     

一类求解线性约束非线性规划问题的可行方向法

Wilson,Han和Powell提出的序列二次规划方法(简称SQP方法)是求解非线性规划问题的一个著名方法,这种方法每次迭代的搜索方向是通过求解一个二次规划子问题得到的,本文受[1]启发,得到二次规划子问题的一个近似解,进而给出了一类求解线性约束非线性规划问题的可行方向法,在约束集合满足正则性的条件下,证明了该算法对五种常用线性搜索方法具有全局收敛性。     

0-1线性规划的连续化及其遗传算法解法

将0-1离散规划通过一个非线性等式约束表示为[0,1]区间上等价的连续变量非线性规划列式.对非线性等式约束的问题进行了两种方法的处理.第一种方法使用乘子法,第二种方法将非线性的等式约束近似为一个非线性的不等式约束,均利用遗传算法程序GENOCOP进行了求解.对多个算例进行了计算,结果表明了该方法的可行性和有效性.     

正交约束优化问题的一阶算法

带有正交约束的矩阵优化问题在材料计算、统计及数据分析等领域中有着广泛的应用.由于正交约束的可行域是Stiefel流形,一直以来流形上的优化方法是求解这一问题的主要方法.近年来,随着实际应用问题所要求的变量规模的扩大,传统的流形优化方法在计算上的劣势显现出来,而一些迭代简单、收敛快的新算法逐渐被提出.通过收缩方法、非收缩可行方法、不可行方法三个类别分别来介绍求解带有正交约束的矩阵优化问题的最新算法.通过分析这些方法的主要特性,以及应用问题的要求,对这类问题算法设计的研究进行了展望.     

半无限规划离散化问题一个两阶段序列二次规划算法

本文研究了求解半无限规划离散化问题(P)的一个新的算法.利用序列二次规划(SQP)两阶段方法和约束指标集的修正技术,提出了求解(P)的一个两阶段SQP算法.算法结构简单,搜索方向的计算成本较低.在适当的条件下,证明了算法具有全局收敛性.数值试验结果表明算法是有效的.推广了文献[4]中求解(P)的算法.     

求解随机凸规划概率约束问题的对偶算法

1引言随机规划中的概率约束问题在工程和管理中有广泛的应用.因为问题中包含非线性的概率约束,它们的求解非常困难.如果目标函数是线性的,问题的求解就比较容易.给出了一个求解随机线性规划概率约束问题的综述.原-对偶算法和切平面算法是比较有效的.在本文中,我们讨论随机凸规划概率约束问题:     

求连续minimax问题整体解的区间算法

1 引 言 Minimax问题是一类重要的数学规划问题,它来源于实际并有极广泛的应用([1],[2]).用区间数学方法求解 minimax问题已取得了一些成果.文[1]对由 C2类函数构成的无约束连续 minimax问题进行了研究,建立了相应的区间算法,文[6]～[11]分别讨论和建立了无约束和不等式约束的离散minimax问题的区间算法.文[12]、[13]则讨论了最坏     

对带有盒约束的二次整数规划的一种线性化方法

本文主要讨论了二次整数规划问题的线性化方法.在目标函数为二次函数的情况下,我们讨论了带有二次约束的整数规划问题的线性化方法,并将文献中对二次0-1问题的研究拓展为对带有盒约束的二次整数规划问题的研究.最终将带有盒约束的二次整数规划问题转化为线性混合本文主要讨论了二次整数规划问题的线性化方法.在目标函数为二次函数的情况下,我们讨论了带有二次约束的整数规划问题的线性化方法,并将文献中对二次0-1问题的研究拓展为对带有盒约束的二次整数规划问题的研究.最终将带有盒约束的二次整数规划问题转化为线性混合0-1整数规划问题,然后利用Ilog-cplex或Excel软件中的规划求解工具进行求解,从而解决原二次整数规划.     

Robustly stable pole-placement based adaptive control of continuous linear systems with multiestimation

This paper deals with the pole-placement type robust adaptive control of continuous linear systems in the presence of bounded noise and a common class of unmodeled dynamics with the use of multiple estimation schemes working in parallel. The multiestimation scheme consisting of the above set of various single estimation schemes is a tool used to minimize the plant identification error by building an estimate which is a convex combination of the estimates at all time. The weighting functions of the individual estimates are provided at each time by a suboptimization scheme for a quadratic loss function of a possibly filtered tracking error and/or control input. The robust stability of the overall adaptive scheme is ensured by an adaptation relative dead zone which takes into account the contribution of the unmodeled dynamics and bounded noise. The basic results are derived for two different estimation strategies which have either a shared regressor with the plant or individual regressors for the input contribution and its relevant time-derivatives. In this second case, the plant input is obtained through a similar convex combination rule as the one used for the estimators in the first approach. An extension of the basic strategies is also pointed out including a combined use of the suboptimization scheme with a supervisor of past measures for the on-line calculation of the estimator weights in the convex combination.     

线性规划的支撑方法（二）

线性规划的支撑方法（二）高学东，武森，李宗元（北京科技大学管理学院，北京１０００８３）６初始支撑可行解的构造６．１初始可行解的构造在研制或设计一种新产品的时候，初始设计往往可以用来帮助构造数学模型，相应的设计向量天虽不一定完全可行，但在某些方面有可能...     

Implementation of gradient methods by tangential discretization

This paper deals with a procedure for implementing iterative methods for nonlinear programming. For constrained problems, we examine the procedure in relation to the gradient-projection method. At each iteration, the domain of suboptimization is replaced by an infinite but discrete set of points, satisfying thetangential properties for the convergence of the algorithm. It should be possible to use this procedure on other iterative methods which proceed by a series of suboptimizations, if the domain of these suboptimizations is of small dimension.The subject of this paper was presented at the Colloque National d'Analyse Numérique, La Grande Motte, France, 1975.The author is indebted to the referees for helpful comments and suggestions.     

Combined location-routing problems: A synthesis and future research directions

Over the past few decades, the concept of integrated logistics systems has emerged as a new management philosophy which aims to increase distribution efficiency. Such a concept recognizes the interdependence among the location of facilities, the allocation of suppliers and customers to the facilities, and the vehicle route structure around depots. As such, it coordinates a broader spectrum of location and routing options available to logistics managers and consequently avoids the suboptimization of distribution solutions. Reflecting the increasing importance of integrated logistics systems, an extensive body of combined location-routing literature has developed in less than 30 years. In this paper, we synthesize the past evolution of location-routing literature and then explore promising research opportunities in incorporation of more realistic aspects, algorithmic design, and model complexity.     

Parallelizing the dual revised simplex method

This paper introduces the design and implementation of two parallel dual simplex solvers for general large scale sparse linear programming problems. One approach, called PAMI, extends a relatively unknown pivoting strategy called suboptimization and exploits parallelism across multiple iterations. The other, called SIP, exploits purely single iteration parallelism by overlapping computational components when possible. Computational results show that the performance of PAMI is superior to that of the leading open-source simplex solver, and that SIP complements PAMI in achieving speedup when PAMI results in slowdown. One of the authors has implemented the techniques underlying PAMI within the FICO Xpress simplex solver and this paper presents computational results demonstrating their value. In developing the first parallel revised simplex solver of general utility, this work represents a significant achievement in computational optimization.     

Mathematical Analysis of Preventive Maintenance Schemes

A new approach is proposed for the maximization of profit by optimal scheduling of machinery. Only one objective function (profit) is used instead of two (availability and cost). The latter approach inevitably resulted in suboptimization. In addition, the "single objective function" approach naturally lends itself to comparisons of efficiency between any preventive maintenance policies. Optimal solutions were found in order to compare the efficiency of the commonly used policies of age and block replacement. Numerical results show that age replacement is always more profitable. Optimal solutions for these two maintenance policies were also found in the specific case where a maintenance repair is superior in quality to a breakdown repair. Finally, the physical law of increasing entropy, applied to the failure rate concept, leads to the conclusion that preventive maintenance should always be considered.     

On a new naturally indexed quick clustering method with a global objective function

A new clustering method is presented which proposes a class of objective functions and an algorithm which sub-optimizes the objective functions over the whole space of partitions. The objective functions have a global nature, encompassing both the cluster contents and the cluster number. However, the accompanying suboptimization algorithm works according to a simple progressive merger scheme. The algorithmic scheme produces in a quite natural way an indexed hierarchy. The hierarchy index is not just tacked on to the method—see Diday and Moreau¹—on the contrary, the algorithm refers directly to its values which measure, depending upon the particular formulation, either the relative affinity or the relative difference of the two clusters merged at a given level of hierarchy. In this way, the scale of hierarchy and hierarchy-wise validity of clusters can easily be established, which is of great importance in analysing unstructured data sets whose generating process is unknown and can only be hypothesized after an initial structure had been established, e.g. owing to clustering, as is the case in pattern recognition—see Kaminuma².     

Comments on: A comparative analysis for determining optimal price and order quantity when a sale increases demand

Although the net present value (NPV) criterion is theoretically the correct approach to developing optimal inventory policies, in the classical EOQ case, the average profit criterion generates solutions that are practically identical to those resulting from the NPV criterion. Nevertheless, a recent paper suggests that, when the demand for a product is price-elastic and a wholesaler offers a one-time-only price discount, use of the average profit criterion may obtain policies that are drastically suboptimal compared to the policies obtained by using the NPV criterion. We show that this suggestion is based on inaccurate models and inconsistent comparisons. Although in cases of large one-time-only discounts, there may be significant differences in the policies and consequences resulting from the two criteria, such large discounts are unrealistic. Furthermore, the larger the discount, the less practicable are the optimal order quantities based on either one of these criteria. Thus, in most real-life situations, the use of the average profit criterion does not result in serious suboptimization. In these situations, what may be important is not whether a retailer uses the NPV criterion or the average profit criterion, but whether the retailer can and does implement the optimal decisions resulting from the use of either criterion.     

Toader型平均与几种经典平均的确界

给出了最佳参数α_1,α_2,α_3,β_1,β_2,β_3∈R,使得双向不等式α_1Q(a,b)+(1-α_1)G(a,b)0且a≠b成立.其中A(a,b)=(a+b)/2,H(a,b)=2ab/(a+b),G(a,b)=(ab)~(1/2),Q(a,b)=((a~2+b~2)/2)~(1/2),C(a,b)=(a~2+b~2)/(a+b),T(a,b)=2/π∫_0~(π/2)(a~2cos~2t+b~2sin~2)~(1/2)tdt分别是两个正数a和b的算术平均,调和平均,几何平均,二次平均,反调和平均和Toader平均.