Randomized extended average block Kaczmarz method for inconsistent tensor equations under t-product
摘要
The Kaczmarz method is widely recognized as a mainstream technique for solving linear systems. To solve large inconsistent tensor equations under tensor t-product, we propose a tensor randomized extended average block Kaczmarz method. We prove theoretically that it converges to the least-squares norm solution of the equation as expected and demonstrate its convergence and performance through numerical experiments. Compared to existing methods, this approach reduces computation time for solving ill-conditioned equations and enhances convergence stability.