Presentation on ICIAM TOKYO (Decided)
Date:
Nonlinear conjugate gradient method for vector optimization on Riemannian manifolds
Kangming Chen（coauthor：Hiroyuki Sato, Ellen Hidemi Fukuda）
Abstract： In this research, we propose a conjugate gradient descent algorithm for vector optimization on Riemannian manifolds. We extend the concepts of Wolfe conditions and Zoutendjik conditions to Riemannian manifolds. The convergence of the proposed method is proved for different choices of the parameter beta, including the Riemannian extension of Fletcher-Reeves, Conjugate Descent, and Dai-Yuan. Numerical experiments are conducted to validate the proposed method.