{"id":"arxiv/2603.11552","url":"https://arxiv.org/abs/2603.11552","tags":["math.NA"],"score":0,"title":"Deep Domain Decomposition Method for Solving the Variational Inequality Problems","domain":"math.NA","source":"arxiv","authors":["Yiyang Wang","Qijia Zhou","Shengyuan Deng","Chenliang Li"],"reviews":0,"subject":"math.NA","summary":"By integrating physics-informed neural network (PINN) techniques with domain decomposition method, a deep domain decomposition method is presented for solving elliptic variational inequality problems. Based on the Ritz variation method, the elliptic variational inequality problem is firstly reformulated as an optimization problem, and then the subproblem in each subdomain is solved by using the Ritz-PINN method, which the parameters in the network are updated by the Adam optimizer, and the residual-adaptive training by introducing a residual-adaptive dataset update strategy to gradually…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-12","reactions":{}}