Relaxed Policy Iteration Algorithm for Nonlinear Zero-Sum Games With Application to H-Infinity Control | Synapse