Pure Mathematics
Vol. 13  No. 05 ( 2023 ), Article ID: 66353 , 4 pages
10.12677/PM.2023.135144

希尔伯特空间上混合变分不等式解的存在性 条件

王欣睿

西南石油大学理学院,四川 成都

收稿日期:2023年4月22日;录用日期:2023年5月24日;发布日期:2023年5月31日

摘要

混合变分不等式是一种重要的变分不等式,它与众多经济学、物理学的问题息息相关。研究变分不等式的一个重要的问题就是变分不等式的解是否存在。本文利用FKKM定理研究了混合变分不等式在各种适当条件下的解的存在性,给出了几个使得MVI解存在的定理。

关键词

混合变分不等式,FKKM定理

Existence Conditions of Solutions for Mixed Variational Inequalities in Hilbert Space

Xinrui Wang

School of Science, Southwest Petroleum University, Chengdu Sichuan

Received: Apr. 22nd, 2023; accepted: May 24th, 2023; published: May 31st, 2023

ABSTRACT

Mixed variational inequality is an important type of variational inequality that is closely related to many problems in economics and physics. An important question in studying variational inequality is whether the solution of the variational inequality exists. This paper uses the FKKM theorem to study the existence of solutions for mixed variational inequality under various appropriate conditions, and provides several theorems that enable the existence of MVI solutions.

Keywords:Mixed Variational Inequalities, FKKM Theorem

Copyright © 2023 by author(s) and Hans Publishers Inc.

This work is licensed under the Creative Commons Attribution International License (CC BY 4.0).

http://creativecommons.org/licenses/by/4.0/

1. 引言

众所周知,变分不等式是研究经济、管理和工程中出现的各种网络均衡问题的有用工具 [1] - [11] 。因此,它引起了数学、物理学、经济学等领域大量研究人员的广泛关注。最初是Lescarret [12] 和Browder [13] 结合众多应用考虑带有非线性项的混合变分不等式,混合变分不等式是一般的变分不等式的一个有用且重要的推广。通常研究变分不等式解的存在性方法有这几种方法:其一,灵活的应用几个经典的大定理,如Brower不动点定理、KKM定理、Ky Fan极大极小原理等,这种方法可以看作是经典不动点理论的一个重要应用;其二,将变分不等式转化为等价的不动点问题,构造迭代算法,然后利用空间的完备性证明迭代点列收敛到变分不等式的解。常用技巧有投影方法、预解方程和预解算子技巧等。对于这方面的研究,可以参阅 [14] [15] 及其参考文献。本文给出了一系列MVI和MMVI的解的存在性及唯一性的定理,利用FKKM定理我们可以得到只要存在一个点使得 { x X : T x , y 0 x + f ( y 0 ) f ( x ) 0 } 为紧集,MVI的解就存在。而如果空间为有限维,在有限维的希尔伯特空间中,有界闭和紧等价,而强制性条件往往可以限制该集合的有界性,此时MVI的解就存在。

2. 准备知识

定义1.1 K为H上的一非空闭凸集, f : K ( , + ] 为一真凸下半连续泛函,

T : H H 连续,混合变分不等式(简记为MVI)为寻找 x K ,使得

T x , y x + f ( y ) f ( x ) 0 y K .

定义1.2 设K是一线性空间X的非空子集,集值映射 M : K 2 X 被称为KKM映

像,如果对于K中的任意有限集 { y 1 , y 2 , , y n }

c o { y 1 , y 2 , , y n } i = 1 n M ( y i )

其中 c o { y 1 , y 2 , , y n } 表示集合 { y 1 , y 2 , , y n } 的凸包。

定理1.1 (FKKM定理)设K是一Hausdorff拓扑线性空间X的非空凸集,

M : K 2 X 是闭值的KKM映像。如果存在一点 y 0 K 使得 M ( y 0 ) 为K中的紧集,那么

y K M ( y )

3. 主要内容

定理2.1 在定义1.1中定义的MVI中,若存在一点 y 0 K 使得

{ x X : T x , y 0 x + f ( y 0 ) f ( x ) 0 }

为K中的紧集。则混合变分不等式有解。

证明:定义一集值映射 M : K 2 X 如下

M ( y ) = { x X : T x , y x + f ( y ) f ( x ) 0 }

首先证明 M ( y ) 为KKM映像。若 M ( y ) 不为KKM映像则存在 y 0 K ,满足

y 0 = i = 1 n λ i y i ,使得 y 0 i = 1 n M ( y i ) 。其中 y 1 , y 2 , , y n K λ i [ 0 , 1 ] i = 1 n λ i = 1

于是对于任意的 i = 1 , 2 , , n

T y 0 , λ i ( y i y 0 ) + λ i f ( y i ) λ i f ( y 0 ) < 0

把n个式子求和可以得到

T y 0 , i = 1 n λ i y i i = 1 n λ i y 0 + i = 1 n λ i f ( y i ) f ( y 0 ) < 0

继而得到

i = 1 n λ i f ( y i ) < f ( y 0 ) = f ( i = 1 n λ i y i ) i = 1 n λ i f ( y i )

故而矛盾,因此 M ( y ) 为KKM映像。

然后我们证明对于任给的 y K M ( y ) 为闭集。 y M ( y ) 故而 M ( y )

{ x n } M ( y ) ,且 x n x * ( n + ) 。则有下面这个式子成立

T x n , y x n + f ( y ) f ( x n ) 0

对这个式子取下极限,并由T的连续性和f的下半连续性可以得到

0 lim inf n + ( T x n , y x n + f ( y ) f ( x n ) ) T x * , y x * + f ( y ) f ( x * )

于是我们知道 x * M ( y ) 。故而 M ( y ) 为闭集。

有条件存在一点 y 0 K 满足

M ( y 0 ) = { x X : T x , y 0 x + f ( y 0 ) f ( x ) 0 }

为K中的紧集。故而由FKKM定理可知 y K M ( y ) 非空,故而MVI有解。

推论2.1 若H为紧空间,则MVI的解存在。

证明:由定理2.1可知对于任意的 y H M ( y ) 为H的闭子集,故而 M ( y ) 为紧集,故而该变分不等式有解。

定理2.2 若H为有限维的希尔伯特空间,若满足存在一点 y 0 K 使得

lim inf x + T x , x y 0 + f ( x ) x = + f ( y 0 ) + 则MVI解存在。且解集为有界闭集。

证明:若 M ( y 0 ) 无界,那么存在一个无界序列 { x n } M ( y 0 ) ,取一个包含于

{ x n } ( B ( 0 , 1 ) ) c 的子序列,不失一般的我们依旧记为 { x n } 。那么对于任意的 n N + ,都有

T x n , y 0 x n + f ( y 0 ) f ( x n ) 0

通过变型我们可以得到

T x n , x n y 0 + f ( x n ) x n f ( y 0 ) x n

n + ,我们可以得到

+ = lim inf n + T x n , x n y 0 + f ( x n ) x n lim inf n + f ( y 0 ) x n 0

矛盾,故而 M ( y 0 ) 有界,由定理2.1我们知道 M ( y 0 ) 为闭集,且H为有限维的希尔伯特空间,故而 M ( y 0 ) 为紧集,由定理2.1知混合变分不等式的解存在。 M ( y 0 ) 有界,易知 y K M ( y ) 有界。故而MVI的解集为有界闭集。

由上面的证明过程易知如下推论

推论2.2 若MVI的解存在,且满足存在一点 y 0 K 使得

lim inf x + T x , x y 0 + f ( x ) x = + f ( y 0 ) +

则MVI的解集为有界闭集。

文章引用

王欣睿. 希尔伯特空间上混合变分不等式解的存在性条件
Existence Conditions of Solutions for Mixed Variational Inequalities in Hilbert Space[J]. 理论数学, 2023, 13(05): 1418-1421. https://doi.org/10.12677/PM.2023.135144

参考文献

  1. 1. Hu, X. and Wang, J. (2006) Solving Pseudomonotone Variational Inequalities and Pseudoconvex Optimization Problems Using the Projection Neural Network. IEEE Transactions on Neural Networks, 17, 1487-1499. https://doi.org/10.1109/TNN.2006.879774

  2. 2. Barbagallo, A., Daniele, P., Lorino, M., Maugeri, A. and Mirabella, C. (2013) Further Results for General Financial Equilibrium Problems via Variational Inequalities. Journal of Mathematical Finance, 3, 33-52. https://doi.org/10.4236/jmf.2013.31003

  3. 3. Kinderlehrer, D. and Stampacchia, G. (1987) An Introduction to Variational Inequalities and Their Applications. In: Classics in Applied Mathematics, Series No. 31, Society for Industrial and Applied Mathematics, Philadelphia.

  4. 4. Oden, J.T. and Kikuchi, N. (1980) Theory of Variational Inequalities with Applications to Problems of Flow through Porous Media. International Journal of Engineering Science, 18, 1173-1284. https://doi.org/10.1016/0020-7225(80)90111-1

  5. 5. Kučera, M. (1977) A New Method for the Obtaining of Ei-genvalues of Variational Inequalities of the Special Type (Preliminary Communication). Commentationes Mathematicae Universitatis Carolinae, 18, 205-210.

  6. 6. Ceng, L.-C. and Yao, J.-C. (2007) An Extragradient-Like Approximation Method for Variational Inequality Problems and Fixed Point Problems. Applied Mathematics and Computation, 190, 205-215. https://doi.org/10.1016/j.amc.2007.01.021

  7. 7. Mignot, F. and Puel, J.P. (1984) Optimal Control in Some Varia-tional Inequalities. SIAM Journal on Control and Optimization, 22, 466-476. https://doi.org/10.1137/0322028

  8. 8. Hu, X. and Wang, J. (2007) A Recurrent Neural Network for Solving a Class of General Variational Inequalities. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 37, 528-539. https://doi.org/10.1109/TSMCB.2006.886166

  9. 9. Verma, R.U. (2001) Projection Methods, Algorithms, and a New System of Nonlinear Variational Inequalities. Computers & Mathematics with Applications, 41, 1025-1031. https://doi.org/10.1016/S0898-1221(00)00336-9

  10. 10. Quittner, P. (1989) Solvability and Multiplicity Results for Variational Inequalities. Commentationes Mathematicae Universitatis Carolinae, 30, 281-302.

  11. 11. Zhang, Y. (2023) Stochastic Variational Inequality Approaches to the Stochastic Generalized Nash Equilibrium with Shared Constraints. Journal of Computational Mathematics, 41, 415-436. https://doi.org/10.4208/jcm.2109-m2020-0099

  12. 12. .Lescarret, C. (1965) Cas d’addition des applications mono-tones maximales dans un espace de Hilbert. Comptes Rendus Hebdomadaires des Séances de l’Académie des Sciences, 261, 1160-1163.

  13. 13. Browder, F.E. (1966) On the Unification of the Calculus of Variations and the Theory of Monotone Nonlinear Operators in Banach Spaces. Proceedings of the National Academy of Sciences of the United States of America, 56, 419-425. https://doi.org/10.1073/pnas.56.2.419

  14. 14. Facchinei, F. and Pang, J.S. (2003) Finite-Dimensional Variational Inequalities and Complementarity Problems, Springer, New York. https://doi.org/10.1007/b97544

  15. 15. Huang, N.-J., Liu, Y.-P., Tang, Y.-Y. and Bai, M.-R. (1999) On the Generalized Set-Valued Strongly Nonlinear Implicit Vari-ational Inequalities. Computers & Mathematics with Applications, 37, 29-36. https://doi.org/10.1016/S0898-1221(99)00123-6

期刊菜单