This release introduces the augmented Lagrangian solver in the DDP solver. The followings are the most important changes:
+ Introduction of Augmented Lagrangian terms and collections (for state and state-input cases)
+ Updating OCP to use augmented Lagrangian collections
+ Introduction of Augmented penalties for AL: two equalities and two inequalities
+ Introduction of Multiplier, MultiplierCollection, LagrangianMetrics, MetricsCollection, and ProblemMetrics
+ Updating approximateIntermediateLQ, approximatePreJumoLQ, and approximateFinalLQ to use MultiplierCollection
+ Introducing Dual solution (final, preJump, intermeiate)
+ Introduction of helper function for initialization and updating metrics, multipliers, and dual solution.
+ Introduction of SolverObserverModule to extract each term’s Multiplier and Metrics from the solver.
+ Introduction of RosSolverObserverModule to publish a term’s Multiplier and Metrics at a specific time on the horizon.
+ Using AL in CartpoleTest and HybridSlqTest