We introduce SILS, a subspace inertial line search algorithm, which is a line search method designed for finding zeros of monotone mappings. At each iteration, a new point is generated in a subspace of the previous points, replacing the one with the largest residual norm. This study analyzes the global convergence and complexity bounds for SILS, including the number of iterations and function evaluations required. Numerical results show that SILS is promising compared to BasicLS with several known derivative-free directions.
View A subspace inertial method for derivative-free nonlinear monotone equations