AQFC2015

Differential Equation Methods for Solving Second-Order Cone Constrained Variational Inequality Problems

 
----------------------------------------------------------------------------------------------------
 
         Department of Systems Engineering and Engineering Management
 
                                 The Chinese University of Hong Kong
 
----------------------------------------------------------------------------------------------------
 
Date: Friday, May 15, 2026, 4:30pm to 5:30pm HKT
 
Venue: ERB513, The Chinese University of Hong Kong
 
Title: Differential Equation Methods for Solving Second-Order Cone Constrained Variational Inequality Problems
 
Speaker: Prof. Juhe Sun, Shenyang Aerospace University
 
 
 
 
Abstract:
 
We propose two efficient approaches for solving the second-order cone constrained variational inequality (SOCCVI): a first-order differential equation (1-ODE) based neural network method and a second-order differential equation (2-ODE) method. For the neural network, we use the Karush–Kuhn–Tucker (KKT) conditions of variational inequality (VI), recasting SOCCVI into an equation system with a smoothing function for metric projection to handle the complementarity condition; we further explore second-order sufficient conditions to achieve exponential stability and prove the non-singularity of the KKT Jacobian under such conditions and constraint nondegeneracy. For the 2-ODE method, we employ a complementary function to reformulate the KKT conditions into a smooth equation system, transform it into an unconstrained optimization problem, and analyze its convergence properties in comparison with the 1-ODE approach. Finally, numerical experiments validate the efficiency of both methods: the neural network outperforms existing ones in stability and convergence rate, while the 2-ODE method serves as a competitive alternative with distinct convergence features.
 
 
 
Biography:
 
Juhe Sun received a Ph.D. degree in Mathematics from Dalian University of Technology, Dalian, China in 2009. From 2009 to 2010, she conducted postdoctoral research at National Taiwan Normal University. She is currently a Professor in the School of Science, Shenyang Aerospace University, Shenyang, China. She has published about 20 papers in refereed journals, including IEEE Transactions and Nonlinear Analysis. She has served as the principal investigator for two projects funded by the National Natural Science Foundation of China (the Tianyuan Special Fund and the Youth Fund), as well as two provincial research projects. Her research interests mainly focus on optimization theory and algorithms.
 
 
 
Everyone is welcome to attend the talk!
 
SEEM-5202 Website: http://seminar.se.cuhk.edu.hk
 
Date: 
Friday, May 15, 2026 - 16:30