A Parameterized Multi-Splitting Iterative Method for Solving the PageRank Problem

In this paper, a new multi-parameter iterative algorithm is proposed to address the PageRank problem based on multi-splitting iteration method. The method solves two linear subsystems at each by splitting coefficient matrix, considering therefore inner and outer find approximate solutions of these subsystems. It can be shown that sequence generated finally converges vector when parameters satisfy certain conditions. Numerical experiments show has better convergence numerical stability than existing algorithms.