A Power Method to Convex-Concave Minmax Optimization Problems with Nonlinear Constraints
- Resource Type
- Conference
- Authors
- Tang, Hao; Chen, Xiao-Feng
- Source
- 2023 International Conference on New Trends in Computational Intelligence (NTCI) New Trends in Computational Intelligence (NTCI), 2023 International Conference on. 1:370-374 Nov, 2023
- Subject
- Computing and Processing
Robotics and Control Systems
Signal Processing and Analysis
Nonlinear equations
Market research
Optimization
Computational intelligence
Min-max optimization problems
Mixed non-linear complementarity problems
Variational inequality
Power penalty methods
Convergence rates
- Language
In this paper, we presents a new power penalty method to convex-concave minimax optimization problems with nonlinear constraints. This method first reformulate the KKT conditions of such optimization problem approximated by a nonlinear equation containing a power penalty term. Then we will show that the solution of the penalty equation converges to that of the mixed complementarity problem at an exponential rate, which is depending on the parameters in the penalty equation, and some numerical results are presented to verify the theoretical results.