A modified Frank-Wolfe algorithm and its convergence properties

ＡＭＯＤＩＦＩＥＤＦＲＡＮＫ－ＷＯＬＦＥＡＬＧＯＲＩＴＨＭＡＮＤＩＴＳＣＯＮＶＥＲＧＥＮＣＥＰＲＯＰＥＲＴＩＥＳＷＵＦＡＮＧ（吴方）ＷＵＳＨＩＱＵＡＮ（吴士泉）（ＩｎｓｔｉｔｕｔｅｏｆＡｐｐｌｉｅｄＭａｔｈｅｍａｔｉｃｓ,ｔｈｅＣｈｉｎｅｓｅＡｃａｄ...     

The Euler scheme and its convergence for impulsive delay differential equations

This paper presents a fixed stepsize Euler scheme for linear impulsive delay differential equations and considers its convergence. We propose a method to take the partition nodes for the Euler scheme. Employing the induction and the technique of inequality, we obtain the order of convergence for Euler scheme. An example is given to illustrate the efficiency of our result.     

On the convergence of solutions of some evolution differential equations

Two conjectures concerning the asymptotic behaviour of solutions of two types of evolution differential equations are presented.     

A modified SLP algorithm and its global convergence

This paper concerns a filter technique and its application to the trust region method for nonlinear programming (NLP) problems. We used our filter trust region algorithm to solve NLP problems with equality and inequality constraints, instead of solving NLP problems with just inequality constraints, as was introduced by Fletcher et al. [R. Fletcher, S. Leyffer, Ph.L. Toint, On the global converge of an SLP-filter algorithm, Report NA/183, Department of Mathematics, Dundee University, Dundee, Scotland, 1999]. We incorporate this filter technique into the traditional trust region method such that the new algorithm possesses nonmonotonicity. Unlike the tradition trust region method, our algorithm performs a nonmonotone filter technique to find a new iteration point if a trial step is not accepted. Under mild conditions, we prove that the algorithm is globally convergent.     

Scheduling flow shops using differential evolution algorithm

This paper describes a novel optimization method based on a differential evolution (exploration) algorithm and its applications to solving non-linear programming problems containing integer and discrete variables. The techniques for handling discrete variables are described as well as the techniques needed to handle boundary constraints. In particular, the application of differential evolution algorithm to minimization of makespan, flowtime and tardiness in a flow shop manufacturing system is given in order to illustrate the capabilities and the practical use of the method. Experiments were carried out to compare results from the differential evolution algorithm and the genetic algorithm, which has a reputation for being very powerful. The results obtained have proven satisfactory in solution quality when compared with genetic algorithm. The novel method requires few control variables, is relatively easy to implement and use, effective, and efficient, which makes it an attractive and widely applicable approach for solving practical engineering problems. Future directions in terms of research and applications are given.     

Eigenvalues of convex processes and convergence properties of differential inclusions

In this paper, we study (real) eigenvalues and eigenvectors of convex processes, and provide conditions for the existence of eigenvectors in a given convex coneK ⁿ . It is established that the maximal eigenvalue ofG(·) inK is expressed by (whereK ⁰ is the polar cone ofK) provided that the minimum is attained in intK ⁰. This result is applied to study the asymptotic behaviour of certain differential inclusions{G(x(t)). We extend some known results for the von Neumann-Gale model to our more general framework. We prove that ifx ₀ is the unique eigenvector corresponding to the maximal eigenvalue ₀ ofG(·) inK, then the nonexistence of solutions of a certain special trigonometric form is necessary and sufficient for every viable solutionx(·) to satisfy^- ₀ ^t x(t)cx ₀ ast for somec0. Our method is to study the family of convex conesW =cl{v–x :xK,vG(x) where is any real number. We characterize the maximal eigenvalue ₀ as the minimal for whichW can be separated fromK.The research was supported in part by a grant from the ministry of science and the Maagara special project for the absorption of new immigrants in the Department of Mathematics at Technion.     

多点收缩混沌优化方法及全局收敛性证明

针对目前混沌优化算法在选取局部搜索空间时的盲目性,提出一种具有自适应调节局部搜索空间能力的多点收缩混沌优化方法.该方法在当前搜索空间搜索时保留多个较好搜索点,之后利用这些点来确定之后的局部搜索空间,以达到对不同的函数和当前搜索空间内已进行搜索次数的自适应效果.给出了该算法以概率1收敛的证明.仿真结果表明该算法有效的提高了混沌优化算法的性能,改善了混沌算法的实用性.     

A novel differential evolution algorithm for binary optimization

Differential evolution (DE) is one of the most powerful stochastic search methods which was introduced originally for continuous optimization. In this sense, it is of low efficiency in dealing with discrete problems. In this paper we try to cover this deficiency through introducing a new version of DE algorithm, particularly designed for binary optimization. It is well-known that in its original form, DE maintains a differential mutation, a crossover and a selection operator for optimizing non-linear continuous functions. Therefore, developing the new binary version of DE algorithm, calls for introducing operators having the major characteristics of the original ones and being respondent to the structure of binary optimization problems. Using a measure of dissimilarity between binary vectors, we propose a differential mutation operator that works in continuous space while its consequence is used in the construction of the complete solution in binary space. This approach essentially enables us to utilize the structural knowledge of the problem through heuristic procedures, during the construction of the new solution. To verify effectiveness of our approach, we choose the uncapacitated facility location problem (UFLP)—one of the most frequently encountered binary optimization problems—and solve benchmark suites collected from OR-Library. Extensive computational experiments are carried out to find out the behavior of our algorithm under various setting of the control parameters and also to measure how well it competes with other state of the art binary optimization algorithms. Beside UFLP, we also investigate the suitably of our approach for optimizing numerical functions. We select a number of well-known functions on which we compare the performance of our approach with different binary optimization algorithms. Results testify that our approach is very efficient and can be regarded as a promising method for solving wide class of binary optimization problems.     

THE SIMPLE BLOCK ITERATIVE ALGORITHM OF AN EXTENDED ENTROPY MODEL AND ITS CONVERGENCE

1.ExtendedEntropyModelEntropymodelsarewidelyusedasaforecastingtechniqueinregionalplanning,see11]and[7].Theydescribeinacertainwaythemostprobablespatialinteractionandcanbeusedtoestimateorigin-destinationtripmatrices.Butthemodelsaretoosimple,soitisdifficulttomeetpracticalrequirements.HallefjordandJo..stenl2]putforwardthegeneralizationofthestandardentropymodels.Onthebasisofthisandaccordingtothedemandofthepredictionofthetripamountfortransportationplanning,wearriveatthefollowing:wherep,μandnarepo…     

On the convergence properties of Hildreth's quadratic programming algorithm

We prove the linear convergence rate of Hildreth's method for quadratic programming, in both its sequential and simulateneous versions. We give bounds on the asymptotic error constant and compare these bounds to those given by Mandel for the cyclic relaxation method for solving linear inequalities.Research of this author was partially supported by CNPq grant No. 301280/86.On leave from the Universidade Federal do Rio de Janeiro, Instituto de Matemática, Rio de Janeiro, R.J. 21.910, Brazil. Research of this author was partially supported by NIH grant HL28438.     

A successive quadratic programming algorithm with global and superlinear convergence properties

This paper presents a successive quadratic programming algorithm for solving general nonlinear programming problems. In order to avoid the Maratos effect, direction-finding subproblems are derived by modifying the second-order approximations to both objective and constraint functions of the problem. We prove that the algorithm possesses global and superlinear convergence properties.This work was supported in part by a Scientific Research Grant-in-Aid from the Ministry of Education, Science and Culture, Japan.     

A novel hybrid immune algorithm and its convergence based on the steepest descent algorithm

This paper proposes a novel hybrid immune algorithm (HIA) that can overcome the typical drawback of the artificial immune algorithm (AIA), which runs slowly and experiences slow convergence. The HIA combines the adaptive AIA based on the steepest descent algorithm. The HIA fully displays global search ability and the global convergence of the immune algorithm. At the same time, it inserts a quasi-descent operator to strengthen its local search ability. A good convergence of the HIA with the quasi-descent idea is shown as well. Numerical experiment results show that the HIA successfully improves running speed and convergence performance.     

一种新的修正SR1更新公式及其算法收敛性

SR1更新公式对比其他的拟牛顿更新公式,会更加简单且每次迭代需要更少的计算量。但是一般SR1更新公式的收敛性质是在一致线性无关这一很强的条件下证明的。基于前人的研究成果,提出了一种新的修正SR1公式,并分别证明了其在一致线性无关和没有一致线性无关这两个条件下的局部收敛性,最后通过数值实验验证了提出的更新公式的有效性,以及所作出假设的合理性。根据实验数据显示,在某些条件下基于所提出更新公式的拟牛顿算法会比基于传统的SR1更新公式的算法收敛效果更好一些。     

Parameter analysis based on stochastic model for differential evolution algorithm

In this paper, a stochastic model is used to describe and analyze the evolution process of differential evolution (DE) for numerical optimization. With the model, it illustrates how the probability distribution of the whole population is changed by mutation, selection and crossover operations. Based on the theoretical analysis, some guidelines about the parameter setting for DE are provided. In addition, numerical simulations are carried out to verify the conclusions drawn from model analysis.     

A penalty function-based differential evolution algorithm for constrained global optimization

We propose a differential evolution-based algorithm for constrained global optimization. Although differential evolution has been used as the underlying global solver, central to our approach is the penalty function that we introduce. The adaptive nature of the penalty function makes the results of the algorithm mostly insensitive to low values of the penalty parameter. We have also demonstrated both empirically and theoretically that the high value of the penalty parameter is detrimental to convergence, specially for functions with multiple local minimizers. Hence, the penalty function can dispense with the penalty parameter. We have extensively tested our penalty function-based DE algorithm on a set of 24 benchmark test problems. Results obtained are compared with those of some recent algorithms.     

A dynamic clustering based differential evolution algorithm for global optimization

A dynamic clustering based differential evolution algorithm (CDE) for global optimization is proposed to improve the performance of the differential evolution (DE) algorithm. With population evolution, CDE algorithm gradually changes from exploring promising areas at the early stages to exploiting solution with high precision at the later stages. Experiments on 28 benchmark problems, including 13 high dimensional functions, show that the new method is able to find near optimal solutions efficiently. Compared with other existing algorithms, CDE improves solution accuracy with less computational effort.     

Adaptive variable space differential evolution algorithm based on population distribution

A novel memetic computing optimization algorithms, i.e. an adaptive variable space differential evolution algorithm (AVSDE), is proposed to improve the global optimization performance. AVSDE guides most individuals search in adaptive variable space (AVS) and employs adaptive mutation and adaptive control parameter. In AVSDE, AVS is determined by population global distribution information, and DE’s operators depend on the local information of the distance and direction. The performance of AVSDE is improved by integrating the global information with the local information. In addition, different mutation strategies are selected according to the evolution stage and random probability to balance AVSDE’s exploration and exploitation abilities, and adaptive control parameter is used to further enhance the performance of AVSDE. 19 scalable benchmark functions are employed to demonstrate the performance of AVSDE. Comparing with two well-tuned conventional DE and several state $-$ of-the $-$ art parameter adaptive DE variants, the whole performance of AVSDE is the best. Finally, two experiments are conducted to analyze the effect of the key parameters on AVSDE’s performance, and the optimal parameters are obtained.     

Information sharing impact of stochastic diffusion search on differential evolution algorithm

This work details the research aimed at applying the powerful resource allocation mechanism deployed in stochastic diffusion search (SDS) to the differential evolution (DE), effectively merging a nature inspired swarm intelligence algorithm with a biologically inspired evolutionary algorithm. The results reported herein suggest that the hybrid algorithm, exploiting information sharing between the population elements, has the potential to improve the optimisation capability of classical DE algorithms. This claim is verified by running several experiments using state-of-the-art benchmarks. Additionally, the significance of the frequency within which SDS introduces communication and information exchange is also investigated.     

The equivalent inclusion technique and its convergence

This paper addresses the convergence characteristics of an iterative solution scheme of the Neumann‐type useful for obtaining homogenized mechanical material properties within an RVE. The analysis is based on the idea of “equivalent inclusions” and, within the context of stress/strain analysis, allows modeling of elastically highly heterogeneous bodies with the aid of discrete Fourier transforms. Within the iterative scheme the proof of convergence depends critically upon the choice of an appropriate, auxiliary stiffness matrix, which also determines the speed of convergence. Mathematically speaking it is based on Banach's fixpoint theorem and only results in necessary convergence conditions. However, for all cases of elastic heterogeneity that are of practical importance convergence can be demonstrated.     

Optimal coordination of directional over-current relays using informative differential evolution algorithm

Growing interconnection in distribution system creates new problem for protection engineers. Particularly the design of overcurrent relay coordination in such system is an independent area of research. With the availability of new artificial based optimization algorithm relay coordination research gain a new momentum. Well established artificial based optimization algorithm such as genetic and particle swam optimization are successfully applied for such applications. This paper discusses the application of informative differential evolution algorithm with self adaptive re-clustering technique for selection of TDS and PSM for optimal coordination of directional overcurrent relays. Both continuous as well as discrete version of informative differential evolution algorithm are used for optimization of relay setting. Proper combination of backup relays for each primary relay are identified by using LINKNET graph theory approach. Coordination of directional overcurrent is developed for 9 bus and IEEE 30 bus distribution systems. The aim of problem is to minimize the total operating time of primary relays and eliminate the miscoordination among the primary and backup relay pairs. Discrete types of settings for electromechanical types of relay are also discussed in this paper. Moreover, the relay coordination problem is modified for providing optimal coordination time interval between 0.2 and 0.8 s among all primary and backup relays pairs. The results are compared with hybrid of genetic algorithm – nonlinear programming and sequential quadratic programming. Digsilient power factory software is used for verification of result.