A decomposition approach to solving the Allup trajectory optimization problem