This module contains actors, i.e., entities that directly calculate actions. Actors are inegrated into controllers (agents).


  • All vectors are treated as of type [n,]

  • All buffers are treated as of type [L, n] where each row is a vector

  • Buffers are updated from bottom to top


Actor(dim_output, dim_input, …[, …])

Class of actors.

ActorCALF(safe_controller, *args[, …])

Actor using Critic As a Lyapunov Function (CALF) to constrain the optimization.

ActorCLF(*args, **kwargs)

ActorCLF is an actor class that aims to optimize the decay of a Control-Lyapunov function (CLF).

ActorMPC(dim_output, dim_input, …[, …])

Model-predictive control (MPC) actor.

ActorProbabilisticEpisodic(dim_output, dim_input)

ActorProbabilisticEpisodicAC(dim_output, …)

ActorRPO(dim_output, dim_input, …[, …])

Running (objective) Plus Optimal (objective) actor.

ActorRQL(dim_output, dim_input, …[, …])

Rollout Q-learning (RQL) actor.

ActorSQL(dim_output, dim_input, …[, …])

Staked Q-learning (SQL) actor.

ActorTabular(dim_world[, predictor, …])

Actor minimizing the sum of the running objective and the optimal (or estimate thereof) objective of the next step.