Local convergence of alternating low‐rank optimization methods with overrelaxation

The local convergence of alternating optimization methods with overrelaxation for low-rank matrix and tensor problems is established. analysis based on the linearization method which takes form an SOR iteration a positive semidefinite Hessian can be studied in corresponding quotient geometry equivalent representations. In case, optimal relaxation parameter accelerating determined from rate standard method. This result relies version Young's theorem 2 × $$ 2\times block systems.