Impacted-region optimization for distributed model predictive control systems with constraints