Global Convergence of the Conjugate Descent Method Under a kind of Inexact Line Search Conditions

1　 IntroductionWe consider the unconstrained optimizatoin problem( P) :minf( x) ,x∈ Rnwhere f:Rn→R is continuously differentiable and its gradient f( x) is denoted by g( x) .The conjugate gradient method is useful for calculating local unconstrained optimizationproblem of differentiable functions of many variables,it has the following form:xk 1 =xk λkdk ( 1 .1 )dk=-gk,k =1-gk βkdk- 1 ,k 2 ( 1 .2 )whereλk is a step-size obtained by a line search,andβk is a parameter,which can beobtaine…     

A Shortest Path Algorithm for Multi—stage Network with Linear Parameter

1　IntroductionThere often existsome network optimization problems with parameters in many real prob-lems.But the effective algorithms are not given because of complexity with parameters.Inthis paper,we propose an effective algorithm for solving the short…     

The Similar Structures and Control Problems of Complex Systems

1 Introduction1 We consider the naturally evolving complex systems, such as biotic and social systems. In their evolving process, they always make up their weaknesses and develop their favorable factors. In such a way, their structures are gradually formed. In the structures of these systems, a feature is noteworthy, i.e., the local(partial) structure is similar to the global(overall). For social systems, for instance, in a country, the management organization of a province is similar to tha…     

Rule Extraction from Trained Artificial Neural Network Based on Genetic Algorithm

This paper discusses how to extract symbolic rules from trained artificial neural network (ANN) in domains involving classification using genetic algorithms (GA). Previous methods based on an exhaustive analysis of network connections and output values have already been demonstrated to be intractable in that the scale-up factor increases with the number of nodes and connections in the network. Some experiments explaining effectiveness of the presented method are given as well.     

A Robust Output—Feedback Controller for a Class of Uncertain Nonlinear Systems

1　IntroductionRecently,the nonlinear output- feedback control problems are solved successfully by usingbackstepping for a class of nonlinear systems. The partialstate feedback control problemsolved in Jiang and Praly,1996 is more general than the output- feedback control problem.However,the assumptions made in the paper are too restrictive which does not only re-quire the existence of c,γ andψ1 to be ascertained a priori,but also require them to beknown for an unknown system. If c,γ can n…     

A Regression Analysis Model Based on Wavelet Networks

1　 IntroductionWavelet analysis is one of the newestsubjects developing quickly in the pastfew decades,with theoretical significance and a wide range of application.Wavelet analysis was notused in the study of statistics until a few years ago.And discussions of nonlinear waveletestimating are focused mainly on two kinds of models. The first one is the modelconsidered by Donoho and his coauthors[1— 3] :yi =f( xi) εi,i =1 ,… ,n =2 J 1 ( 1 )wherexi=in∈ [0 ,1 ] ,ε1 ,… ,εnare independente…     

A Knowledge Tree Model for Managing Organzational Knowlege

According to the relation of organizational knowledge,this paper analyzes the structure of organizational knowledge at first.A concept of knowledge tree is introduced and a process model of knowldege management is described by a knowledge tree,In this paper a defintion of value of knowlege is given and the life cylce of a knowledge tree is analyzed.Finally,some principles for knowledge management are presented.     

State Prediction of Chantic System Based on ANN Model

The choice of time delay and embedding dimension is very important to the phase space reconstruction of any chaotic time series.In this paper,we determine optimal time delay by computing autocorrelation function of time series,Optimal embedding dimension is given by means of the relation between embedding dimension and correlation dimension of chaotic time series,based on the methods above,we choose ANN model to appoximate the given true system.At the same time,a new algorithm is applied to determine the network Weights.At the end of this paper,the theory above is demonstrated through the research of time series generated by Logistic map.