tigercontrol.models package¶
control¶
tigercontrol.models.control.ControlModel() |
Description: class for implementing algorithms with enforced modularity |
tigercontrol.models.control.KalmanFilter() |
Description: Kalman Filter adjusts measurements of a signal based on prior states and knowledge of intrinsic equations of the system. |
tigercontrol.models.control.ODEShootingMethod() |
Description: Implements the shooting method to solve second order boundary value problems with conditions y(0) = a and y(L) = b. |
tigercontrol.models.control.LQR() |
Description: Computes optimal set of actions using the Linear Quadratic Regulator algorithm. |
tigercontrol.models.control.MPPI() |
Description: Implements Model Predictive Path Integral Control to compute optimal control sequence. |
tigercontrol.models.control.CartPoleNN() |
Description: Simple multi-layer perceptron policy, no internal state |
tigercontrol.models.control.ILQR() |
Description: Computes optimal set of actions using the Linear Quadratic Regulator algorithm. |
time_series¶
tigercontrol.models.time_series.TimeSeriesModel() |
Description: class for implementing algorithms with enforced modularity |
tigercontrol.models.time_series.AutoRegressor() |
Description: Implements the equivalent of an AR(p) model - predicts a linear combination of the previous p observed values in a time-series |
tigercontrol.models.time_series.LastValue() |
Description: Predicts the last value in the time series, i.e. |
tigercontrol.models.time_series.PredictZero() |
Description: Predicts the next value in the time series to be 0, i.e. |
tigercontrol.models.time_series.RNN() |
Description: Produces outputs from a randomly initialized recurrent neural network. |
tigercontrol.models.time_series.LSTM() |
Description: Produces outputs from a randomly initialized LSTM neural network. |
tigercontrol.models.time_series.LeastSquares() |
Description: Implements online least squares. |
optimizers¶
tigercontrol.models.optimizers.Optimizer([…]) |
Description: Core class for model optimizers |
tigercontrol.models.optimizers.Adagrad([…]) |
Description: Ordinary Gradient Descent optimizer. |
tigercontrol.models.optimizers.Adam([pred, …]) |
Description: Ordinary Gradient Descent optimizer. |
tigercontrol.models.optimizers.ONS([pred, …]) |
Online newton step algorithm. |
tigercontrol.models.optimizers.SGD([pred, …]) |
Description: Stochastic Gradient Descent optimizer. |
tigercontrol.models.optimizers.OGD([pred, …]) |
Description: Ordinary Gradient Descent optimizer. |
tigercontrol.models.optimizers.mse(y_pred, …) |
Description: mean-square-error loss :param y_pred: value predicted by model :param y_true: ground truth value :param eps: some scalar |
tigercontrol.models.optimizers.cross_entropy(…) |
Description: cross entropy loss, y_pred is equivalent to logits and y_true to labels :param y_pred: value predicted by model :param y_true: ground truth value :param eps: some scalar |
boosting¶
tigercontrol.models.boosting.SimpleBoost() |
Description: Implements the equivalent of an AR(p) model - predicts a linear combination of the previous p observed values in a time-series |