一个求解二阶锥变分不等式问题的神经网络
2023-04-29刘怡彤穆学文
刘怡彤 穆学文
本文提出了一个神经网络算法,以求解二阶锥变分不等式 (SOCCVI) 问题. 该算法利用一个光滑化Fischer-Burmeister(FB)函数处理问题对应的KKT条件,将其转化为一个无约束优化问题. 利用Lyapunov方法本文证明,在给定的条件下,该神经网络Lyapunov稳定,渐近稳定且指数稳定.数值模拟验证了该神经网络的运算效果.
神经网络; 二阶锥; Fischer-Burmeister函数; Lyapunov稳定
O224A2023.011002
A neural network for solving the second-order cone constrained variational inequality problems
LIU Yi-Tong, MU Xue-Wen
(School of Mathematics and Statistics, Xidian University, Xian 710126, China)
A neural network is proposed to solve the second-order cone constrained variational inequality (SOCCVI) problems. In this method, a smoothed Fischer-Burmeister (FB) function is used to deal with the KKT conditions corresponding to the problem, and then the KKT conditions are further transformed to an unconstrained optimization problem. The Lyapunov method is applied to show the Lyapunov stability, asymptotic stability and exponential stability of the neural network under given conditions. The effectiveness of the neural network is verified by numerical experiment.
Neural network; Second-order cone; Fischer-Burmeister function; Lyapunov stability
参考文献:
[1] 程欢, 穆学文, 宋琦悦. 一种新的求解圆锥规划的非内点算法 [J]. 四川大学学报: 自然科学版, 2019, 56: 203.
[2] Yang X, Wang H, Liu K, et al. Minimax and WLS designs of digital FIR filters using SOCP for aliasing errors reduction in BI-DAC [J]. IEEE Access, 2019, 7: 11722.
[3] Kanno Y, Yamada H. A note on truss topology optimization under self-weight load: mixed-integer second-order cone programming approach [J]. Struct Multidiscip O, 2017, 56: 221.
[4] Bergounioux M, Piffet L. A second-order model for image denoising [J]. Set-Valued Var Anal, 2010, 18: 277.
[5] Hopfield J J, Tank D W. Neural computation of decision in optimization problems [J]. Biol Cybern, 1985, 52: 141.
[6] Zhang H, Wang Z, Liu D. A comprehensive review of stability analysis of continuous-time recurrent neural networks [J]. IEEE T Neur Net Lear, 2014, 25: 1229.
[7] Jin L, Li S, Hu B, et al. A survey on projection neural networks and their applications[J]. Appl Soft Comput, 2019, 76: 533.
[8] Sun J, Chen J, Co C. Neural networks for solving second-order cone constrained variational inequality problem [J]. Comput Optim Appl, 2012, 51: 623.
[9] Miao X, Chen J, Ko C. A smoothed NR neural network for solving nonlinear convex programs with second-order cone constraints [J]. Inform Sciences, 2014, 268: 255.
[10] Nazemi A, Sabeghi A. A novel gradient-based neural network for solving convex second-order cone constrained variational inequality problems [J]. J Comput Appl Math, 2019, 347: 343.
[11] Nazemi A, Sabeghi A. A new neural network framework for solving convex second-order cone constrained variational inequality problems with an application in multi-finger robot hands [J]. J Exp Theor Artif In, 2020, 32: 181.
[12] Sun J, Wu X, Saheya B, et al. Neural Network for Solving SOCQP and SOCCVI based on two discrete-type classes of SOC complementarity functions [J]. Math Probl Eng, 2019, 2019: 1.
[13] Sun J, Fu W, Alcantara J, et al. A neural network based on the metric projector for solving SOCCVI problem [J]. IEEE T Neur Net Lear, 2021, 32: 2886.
[14] Fukushima M, Luo Z, Tseng P. Smoothing functions for second-order-cone complementarity problems [J]. SIAM J Optimiz, 2002, 12: 436.
[15] Sun D, Sun J. Strong semismoothness of the fischer-burmeister SDC and SOC complementarity functions [J]. Math Program, 2005, 103: 575.
[16] Chi X, Liu S. A one-step smoothing Newton method for second-order cone programming [J]. J Comput Appl Math, 2009, 223: 114.
[17] Sun J, Zhang L. A globally convergent method based on Fischer-Burmeister operators for solving second-order cone constrained variational inequality problems [J]. Comput Math Appl, 2009, 58: 1936.