.png){kind=icon}
.png){kind=icon}
Return
An Operator Splitting-Based Interior-Point Method for Conic Linear Programming
.png){kind=icon}
PRE
AI
DOI：10.1007/s40305-025-00635-7.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We introduce the Douglas-Rachford splitting interior-point method (DRSIP), a novel algorithm that combines operator splitting with second-order approaches to solve conic linear programming problems. Our method originates from the fixed-point mapping derived from the Douglas-Rachford splitting method, which transforms the barrier penalized conic programming problem into a series of nonlinear equations. We apply the Newton-type method with a path-following scheme for acceleration. We prove the global convergence of DRSIP and establish its local quadratic convergence under the strict complementarity condition. Our numerical results showcase the algorithm's robustness, scalability, and adaptability, positioning DRSIP as a versatile and effective solution for large-scale conic programming challenges.
Keywords:
Conic programming
Interior-point methods
Douglas-Rachford splitting
Path following
Journal
J
IF:
1.1
Papers:
64
Citations:
487
Organization
No organization information available