摘要
本文首先给出了混合约束条件下拟可微函数的Fritz John条件,其次推出了关于K-凸化集性质的三个定理,最后基于Fejer映射的基本原理,构造出了一类非光滑半无限规划问题的Fejer算法并证明其收敛性。
第2章在Polyakova正则和Shapiro正则的定义下,给出了混合约束条件下的拟可微函数的Fritz John条件。
第3章首先给出了两种凸化集的定义。针对Jeyakumar和Luc凸化集,定义了正则凸化集,构造了唯一极小正则凸化集和极小凸化集,并给出了凸化集的极值条件。最后,给出了一般Banach空间算子的K-凸化集的基本定义,推出了关于算子K-凸化集性质的三个定理。
第4章首先给出了Fejer映射的基本性质,然后基于Fejer映射的基本原理给出了一类非光滑半无限规划问题的Fejer算法并证明其收敛性。
Firstly, Fritz John conditions for quasi-differentiable functiongs with hybrid constraints are proposed in this paper. Sencondly, three theorems on properties of K'-convexificator of operator are presented. Lastly, based on basic principle of Fejer mapping, we construct Fejer Algorithm for solving a class of nonsmooth semi-infinite programming problems and prove its convergence.In Chapter 2, Fritz John conditions for quasi-differentiable functions with hybrid constraints are presented from the definitions of Polyakova regularity and Shapiro regularity.In Chapter 3,we first introduce two definitions of convexificator. Contraposing Jeyaku-mar and Luc convexificators,we define a regular convexificator. Then we construct unique minimum regular convexificator and minimum convexificator , moreover,some conditions for convexificator are given. At the last of the chapter, the basic definition of K'-convexificator of operator in general Banach space is investigated, and three theorems on properties of K'-covexificatoe of operator are presented.In Chaper 4,we first give some fundamental properties of Fejer mapping. And the based on the fundamental principle of Fejer mapping, Fejer Algorithm for solving a clars of nonsmooth semi-infinite programming problems is presented and prove its convergence.
引文
[1] Rockafellar R T. Convex Analysis. Princeton: Princeton University Press, 1970.
[2] Demyanov V F. The rise of nonsmooth analysis its main tools. Cybernetics and Systems Analysis, 2002, 38(4): 527-547.
[3] Demyanov V F, Dixon. Quasidifferentiable Calculus. Math.Prog, 1986, Study29:.
[4] Demyanov V F, Rubinov A M. Quasidifferentiable Calculus. Publications Division, New York: Optimization Saftware Inc, 1986.
[5] Demyanov V F, Rubinov A M. Nondifferentiable Optimization. Publications Division,Moskow: Physical Mathematical Inc, 1981.
[6] Clarke F H. Optimization and Nonsmooth Analysis. New York: Wiley, 1983.
[7] Demyanov V F, Rubinov A M. Constructive Nonsmooth Analysis. Frankfurt a/M: Verlag Peter Lang, 1995.
[8] 王炜,夏尊铨.U-拉格朗日函数的最优解集.辽宁师范大学学报(自然科学版),2002,25(4):346-348.
[9] 王炜,夏尊铨,庞丽萍.UV-理论在一类半无限最小化问题的应用.运筹学学报,2004,8(3):29-38.
[10] 孟凡文,高岩,夏尊铨.极大值函数的二阶方向导数.大连理工大学学报,1998,38(6):621-624.
[11] 陈志平,徐成贤.极小拟可微函数的二阶方法.工程数学学报,1995,12(2):84-88.
[12] 王先甲,王秋庭,罗荣桂.拟可微函数拟微分的唯一性.武汉工业大学学报,1995,17(2):23-27.
[13] 卢华.等价拟微分集类的有界性.贵州工学院学报,1992,21(1):65-70.
[14] 卢华.Banach空间中的局部Lispcitz函数与拟可微函数.贵州工学院学报,1994,23(1):55-59.
[15] 王希云,张可村.拟可微B-preinvex规划的最优性条件.太原重型机械学院学报,1996,17(4):277-282.
[16] 谭忠富.拟可微函数的二阶极值条件.东北重型机械学院学报,1991,15(3):269-273.
[17] 姚静.拟可微函数最优化问题的极小化方法.吉林大学自然科学学报,2001,(3):7-10.
[18] 张立卫,张鑫.求解拟可微方程组的非精确牛顿法.经济数学,2001,18(1):74-81.
[19] 高岩.两个凸多面体的差及几种次微分之间的关系.运筹学学报,2000,4(3):57-66.
[20] 殷洪友,张可村.一类拟可微优化问题的最优性条件.工程数学学报,1996,13(3):47-52.
[21] 殷洪友,张可村.具有等式约束的拟可微规划的Fritz John条件极其充分条件.高校应用数学学报,1998,13(1):99-104.
[22] 高岩.拟可微函数的弱Fritz John条件.运筹学杂志,1990,9(2):43-44.
[23] 高岩.具有混合约束条件的拟可微函数的Fritz john条件,高校应用数学学报(A辑),1994,9(3):329-334.
[24] 高岩,邓明荣.一类不可微优化的一阶最优性必要条件.东北重型机械学院学报,1996,20(1):58-61.
[25] 高岩.具有等式与不等式约束条件拟可微优化的Lagrange乘子型最优性条件.运筹学学报,1999,3(4):47-54.
[26] 夏凌青.ε-拟可微规划的Fritz John型条件和Kuhn-Tucker型条件.焦作工学院学报,1996,15(2):87-92.
[27] Shapiro A. On optimality conditions in quasidifferentiable optimization. SIAM J Control and Optimization, 1984, 22(4): 610-617.
[28] Polayakova L N. On the minimization of quasidifferentiable function subject to equalitytype quasidifferentiable constrains. Math.prog, 1986, Study(29): 44-55.
[29] Demyanov V F. Quasidifferentiable functions: necessary conditions and descent directions.Math.Prog, 1986, Study(29): 44-55.
[30] Eppler K, Luderer B. The Lagrange principle and quasidifferential calculus. TU Karl-Marx-Stadt, 1987, 29: 187-192.
[31] Xia Z Q. Generalized Fritz John necessary conditions for quasidifferentiable minimization.Optimization and Computingin (Approximation), 1990, Ed. by Law A G, Wang, C L North-Holland: 315-327.
[32] Gao Y. Demyanov difference of two sets and optimality conditions of Lagrange multiplier type for constrained quasidifferentiable optimization. Journal of optimization theory and applications, 2000, 104(2): 377-394.
[33] Demyanov V F, Jeyakumar V. Hunting for a Smaller Convex Subdifferential. Journal of Global Optimization ,1997, 10: 305-326.
[34] Uderzo A. On the Convexification and the Subdifferentiation of Operators in Banach Spaces. Vestnik Perterburskogo Universiteta seriyal mathematical i Astornomiya, 1995, 15: 47-52.
[35] Jeyakumar V, Luc D T. Nonsmooth Calculus, Minimality, and Monotonicity of Conuexifactors. Journal of Optimization Theory and Applications, 1999, 101: 599-621.
[36] Utta J, Chandra S. Convexifactors,generalized convexity, and optimality conditions. Journal of Optimization Theopy and Applications, 2002, 113(1): 41-64.
[37] Crave B D ,Ralph D, Glover B M. Small convex-valued subdifferentials in mathema- tical programming. Optimization, 1995, 33: 1-21.
[38] Demyanov V F. On relationship between the Clarke subdifferential and the quasidifferential. Vestnik of Leningrad University , 1980, 13: 18-24.
[39] Ioffe A D. A Lagrange multiplier rule with small convex-valued subdifferentials for non-smooth problems of mathematical programming involving equality and non-functional constraints.Math. Prog., 1993, 58: 137-145.
[40] Jeyakumar V. On optimality conditions in nonsmooth inequality constrained minimization. Numer. Funct.Anal.and Optim.,1987, 9(5,6): 536-546.
[41] Kuntz L, Scholtes S. Constraint qualifications in quasidifferentiable optimization. Math. Prog., 1993, 60: 339-347.
[42] Pschenichniy B N. Convex Analysis and Extremal Problems. Moscow: Nauka, 1980.
[43] Rubinov A M, Akhundov I S. Difference of compact sets in the sense of Demyanov and its application to nonsmooth analysis. Optimization, 1992, 23: 179-188.
[44] Jeyakumar V, Luc D T. Approximate Jacobian Matrices for Continuous Maps and Cl- Optimization. SIAM Journal on Control and Optimization, 1998, 36, 1815-1832.
[45] Jeyakumar V, Luc D T. Schaible S. Characterizations of Generalized Monotone Non- smooth Continuous Maps Using Approximate Jacobians. Journal of Convex Analysis, 1998, 5: 119-132 .
[46] Morduchovich B S, Shao Y. On Nonconvex Subdifferential Calculus in Banach Spaces. Journal of Convex Analysis, 1995, 2: 211-228.
[47] Treiman J S. The Linear Nonconvex Generalized Gradient and Lagrange Multipliers. SIAM Journal on Optimization, 1995, 5: 670-680.
[48] Wang X, Jeyakumar V. A Sharp Lagrangian Multiplier Rule for Nonsmooth Mathematical Programming Problems Inuoluing Equality Constraints. SIAM Journal on Optimization, 2000, 10: 1136-1148.
[49] Coop I D, Wastom G A. A projected lagragian algorithms for semi-infinite programming . Math. Prog, 1985,(32): 337-356.
[50] Conn A R, Gould I M. An exact penalty function for semi-infinite programming . Math. Prog.,1987,(37):19-40.
[51] Tichatschke R, Nebeling V. A cutting plane algorithm for silving quadratic semi-infinite programs. Optimization,1988, (19): 803-817.
[52] Huth M, Tichatschke R. A bybrid method for solving semi-infinite programming. Methods of Operations Research, 1989, (68): 79-90.
[53] Motzkin T S, Schoenberg J J. The relaxation method for linear inequalities. Canad. J. Math., 1954. Vol.6, No3: 393-404.
[54] Kaplan A A,Tichatschke R A. Regularized penalty method for solving convex semi-infinite programs. Optimization, 1992, (26): 215-228.
[55] 万仲平.非线性半无限规划的一种新的逼近精确罚函数法.应用数学,1996(增刊), 80-83.
[56] Agmon S. The relaxation method for linear inequalities. Ibid. 382-392.
[57] Eremin I I,Mazurov V G. Nonstationary process of mathematical programming. Science Press, 1979. (in Russian)
[58] Patsko V S. Fejer processes for infinite systems of convex inequalities . Izv.Ural.Gos.Univ. Mat.Mekh., 2002, 4(22): 129-147.(in Russian)