国产日韩欧美一区二区三区三州_亚洲少妇熟女av_久久久久亚洲av国产精品_波多野结衣网站一区二区_亚洲欧美色片在线91_国产亚洲精品精品国产优播av_日本一区二区三区波多野结衣 _久久国产av不卡

?

非可微r-不變凸函數(shù)的η-鞍點條件

2023-12-14 19:16袁靜李向有劉文艷
關鍵詞:鞍點

袁靜 李向有 劉文艷

摘 要:利用η-逼近法,定義了η-鞍點和η-Lagrange函數(shù)。研究了一類包含r-不變凸函數(shù)的非線性數(shù)學規(guī)劃問題的鞍點條件,得到了η-近似優(yōu)化問題下的η-鞍點最優(yōu)性準則和原規(guī)劃的最優(yōu)解與η-近似優(yōu)化問題下的η-Lagrange鞍點的等價性,用新的方法推廣了相關鞍點結(jié)論。

關鍵詞:η逼近方法;η-鞍點;r-不變凸函數(shù);η-Lagrange函數(shù)

中圖分類號:O221.6;O224

文獻標志碼:A

近年來,利用相關的向量優(yōu)化問題設計新的方法來解決原有的多目標數(shù)學規(guī)劃問題及其對偶問題受到了廣泛的關注。ANTCZAK[1,2]提出了一種改進目標函數(shù)的新方法,分別用于求解包含不變凸函數(shù)的可微多目標優(yōu)化問題和非線性標量數(shù)學規(guī)劃問題,通過對任意但固定的可行點x處修改給定數(shù)學規(guī)劃問題的目標函數(shù)和約束函數(shù),從而構(gòu)造出η-近似優(yōu)化問題。進一步建立了它與原非凸數(shù)學規(guī)劃問題的等價性。在文獻 [3-6]中基于對不同的目標規(guī)劃問題中通過修改原規(guī)劃問題的目標函數(shù)和約束函數(shù),構(gòu)造得到相關的η-近似優(yōu)化問題,建立其原規(guī)劃問題和η-近似優(yōu)化問題之間的等價性,得到相關的最優(yōu)性條件、對偶條件和鞍點。

在數(shù)學規(guī)劃問題中拉格朗日乘子及其拉格朗日函數(shù)的鞍點已經(jīng)被廣泛研究(如文獻[7-9])。這一結(jié)果在最優(yōu)化理論和經(jīng)濟學中具有重要作用。同時,凸性的假設也是十分重要的,但其具有一定的局限性。為了放寬充分最優(yōu)性定理中的凸性假設,HANSON[10]推廣了凸函數(shù)即為不變凸函數(shù)。后來,ANTCZAK[11]引入了一類新的(非凸)可微函數(shù),并將其稱為關于η的r-不變凸。 BHATIA[12]在多目標規(guī)劃問題中討論了其最優(yōu)性和鞍點準則。HO[13]在多目標分式規(guī)劃問題中討論了r-不變凸函數(shù)的鞍點準則。ANTCZAK[11,14]證明了包含r-不變凸函數(shù)的約束優(yōu)化問題的充分最優(yōu)性條件和Wolfe對偶性條件,給出了關于另一種類型的η-不變性的條件。

本文進一步討論了文獻[2]中的方法,解決一類非線性數(shù)學規(guī)劃問題的r-不變凸函數(shù)關于η相同函數(shù)的問題。通過研究文獻[11]中提出的η-近似方法,在用該方法構(gòu)造的η-近似優(yōu)化問題中,引入了一個所謂η-鞍點的定義和的η-Lagrange函數(shù)的定義。進一步證明了η-近似優(yōu)化問題中η-Lagrange函數(shù)的一個η-鞍點與原數(shù)學規(guī)劃問題的一個最優(yōu)解之間的等價性。

參考文獻:

ANTCZAK T. A new approach to multiobjective programming with a modified objective function[J]. Global Optimization, 2003, 27: 485-495.

[2] ANTCZAK T. An η-approximation approach or nonlinear mathematical programming problems involving invex functions[J]. Num. Functional Anal. Optimization, 2004, 25(5&6): 423-438.

[3] ANTCZAK T. Saddle points criteria in nondifferentiable multiobjective programming with V-invex functions via an η-approximation method[J]. Computers and Mathematics with Applications, 2010, 60: 2689-2700.

[4] ANTCZAK T. Saddle point criteria and duality results in multiobjective programming problems via an η-approximation method[J]. ANZIAMJ, 2005, 47: 155-172.

[5] 閆春雷. 不變凸多目標規(guī)劃對偶性的逼近方法[J].青島大學學報(自然科學版), 2011, 24(3): 1-6.

[6] ANTCZAK T. Saddle points criteria via a second order η-approximation approach for nonlinear mathematical programming problems involving second order invex functions[J]. Kybernetika, 2011, 47(2): 222-240.

[7] 趙勇, 彭再云, 徐先兵, 等.半B-(p,r)-(預)不變凸函數(shù)與多目標分式規(guī)劃問題的鞍點[J]. 重慶師范大學學報(自然科學版), 2012, 29(1): 18-26.

[8] 楊玉紅. 非光滑半無限多目標優(yōu)化問題的Lagrange鞍點準則[J]. 應用數(shù)學學報, 2018, 41(1): 14-26.

[9] 李鈺, 嚴建軍, 李江榮. 具有廣義凸性的一類半無限向量分式規(guī)劃的鞍點準則[J]. 貴州大學學報(自然科學版), 2015, 32(5): 1-4.

[10]HANSON M A. On sufficiency of the Kuhn-Tucker conditions[J]. Math Anal Appl, 1981, 80: 545-550.

[11]ANTCZAK T. A new method of solving nonlinear mathematical programming problems involving r-invex functions[J]. Math Anal Appl, 2005, 311: 313-323.

[12]BHATIA G. Optimality and mixed saddle point criteria in multiobjective optimization[J]. J Math Anal Appl, 2008,342(1):135-145.

[13]HO S C. Saddle point criteria in multiobjective fractional programming involving exponential invexity[J]. Bull Malays Math Sci Soc, 2018,41:1923-1934.

[14]ANTCZAK T. An η-approximation approach to duality in mathematical programming problems involving r-invex functions[J]. Math Anal Appl, 2006, 315: 555-567.

[15]CLARKE F H. Optimization and nonsmooth analysis[M]. New York: Wiley-Interscience, 1983.

(責任編輯:于慧梅)

η-saddle Point Condition for Non-differentiable r-invexity Functions

YUAN Jing, LI Xiangyou*, LIU Wenyan

(College of Mathematics and Computer Science,Yan’an University,Yan’an 716000,China)

Abstract:

This paper uses the η-approximation method to define the η-saddle point and η-Lagrange functions. The saddle point conditions of a class of nonlinear mathematical programming problems including r-invex functions are studied, and the η-saddle point optimality criterion under the η-approximation optimization problem and the optimal solution of the original program and the equivalence of the η-Lagrange saddle point under the η-approximation optimization problem are obtained, and the relevant saddle point conclusion is generalized by the new method.

Key words:

η-approximation method; η-saddle point; r-invariant convex function; η-Lagrange function

猜你喜歡
鞍點
(F,α,ρ,d)-凸多目標分式規(guī)劃的鞍點準則
求解無約束函數(shù)局部鞍點的數(shù)值算法
實值函數(shù)近似鞍點集的連續(xù)性
一種廣義松弛正定反預處理求解非Hermitian鞍點問題
鞍點在FC-度量空間中新不動點定理中的運用
貿(mào)易戰(zhàn)背景下的中美貿(mào)易研究
含有二階冪零鞍點的雙同宿環(huán)附近的極限環(huán)分支
SKT不變凸非線性規(guī)劃的鞍點特征研究
一種改進的基于二階統(tǒng)計量的盲源抽取算法*
改進的復制動態(tài)方程及其穩(wěn)定性分析