A truncated aggregate smoothing Newton method for minimax problems (Q979271)

By combining a truncated aggregate technique with a stabilized Newton method, the authors propose an algorithm to solve finite minimax problems. At each iteration, the number of gradient and Hessian calculations is reduced by aggregating a small subset of components in the max-function. Some truncating criteria are given for adaptively updating the subset to guarantee the global convergence and locally quadratic convergence. Some numerical results are presented to show the effectiveness of the proposed algorithm compared to three other aggregate smoothing algorithms.